Technical Basis Document No. 10: Unsaturated Zone Transport Revision 1 May 2004 1.1 OVERVIEW This technical basis document provides a summary of the conceptual understanding of the general issues affecting transport of radioactive solute and colloidal species through the unsaturated zone of Yucca Mountain and an evaluation of the relative importance of controlling factors and conditions. The postclosure performance of the repository is determined in part by the ability of radionuclides escaping from the waste packages to reach the water table. This document is one in a series that is being prepared for each component of the repository system relevant to predicting its likely postclosure performance. The relationship of radionuclide transport to the other components is illustrated in Figure 1-1. Radionuclide transport is affected by the processes that will be described in Technical Basis Document No. 1: Climate and Infiltration, Technical Basis Document No. 2: Unsaturated Zone Flow, Technical Basis Document No. 3: Water Seeping into Drifts, CSNF Waste Form Degradation: Summary Abstraction (BSC 2003a, Section 4), and Technical Basis Document No. 9: Engineered Barrier System Transport. The amount and rate of radionuclide arrival at the water table after potential releases from the repository has a direct impact on downstream components of the Yucca Mountain system (i.e., on transport in the saturated zone underneath Yucca Mountain and possible later appearance in the biosphere). The information presented in this document, along with the associated references, provides a framework for the radionuclide-transport-related studies supporting the ongoing development of the postclosure safety analysis to be included in the license application (LA). 1.2 PURPOSE AND SCOPE The purpose of this document is to provide an overview of the current understanding of radionuclide transport through the unsaturated zone to the saturated zone, including a summary description of transport modeling activities. Transport modeling and analysis are performed to evaluate the attributes of the unsaturated zone as a natural barrier. This document presents the technical basis for addressing transport-related Key Technical Issue (KTI) agreements between the U.S. Nuclear Regulatory Commission (NRC) and the U.S. Department of Energy (DOE). Detailed responses to these transport-related KTI agreements are provided in Appendices A to C (Table 1-1). 1. INTRODUCTION 1-1 May 2004 No. 10: Unsaturated Zone Transport Figure 1-1. Components of the Postclosure Technical Basis for the License Application Table 1-1. Transport-Related Key Technical Issues Addressed in Appendices KTI ID RT 1.02 RT 3.01 RT 3.04 RT 3.10 TSPAI 3.28 TSPAI 3.29 No. 10: Unsaturated Zone Transport Revision 1 model Appendix A B B A C C Short Description Provide analog radionuclide data from the tracer tests for Calico Hills at Busted Butte Importance of transport through fault zones below the repository Relative importance of hydrogeologic units beneath the repository Provide data from analog tracers used at Busted Butte Confidence in the active-fracture continuum concept in the transport Integration of the active fracture model with matrix diffusion in the transport model May 2004 1-2 Revision 1 Radionuclide transport processes are qualitatively described in Section 2. These processes include solute advection and dispersion, colloidal transport, and retardation mechanisms, such as sorption and matrix diffusion. The properties of the radionuclide species are also discussed, including radioactive daughter products and diffusion coefficients. The relationship of unsaturated zone radionuclide transport to other components of the repository system and radionuclide transport processes is described in Section 3. Section 4 contains a discussion of laboratory data and field tests that provide parameters for radionuclide transport simulation and evidence supporting the conceptual model for unsaturated zone transport. Section 5 covers the development of models for unsaturated zone transport, including descriptions of how the other system components (e.g., the effects of climate, the unsaturated zone flow) are incorporated into the radionuclide transport model. This section also discusses the conceptualization and simplifications used in model development, as well as transport model calibration and confidence building. The predictive simulations are summarized in Section 6. The consideration of model uncertainties is also discussed. The transport abstraction for total system performance assessment (TSPA) is presented in Section 7. Summary and conclusions are presented in Section 8. 1.3 NOTE REGARDING THE STATUS OF SUPPORTING TECHNICAL INFORMATION This document was prepared using the most current information available at the time of its development. This technical basis document and its appendices providing Key Technical Issue agreement responses were prepared using preliminary or draft information and reflect the status of the Yucca Mountain Project’s scientific and design bases at the time of submittal. In some cases, this involved the use of draft analysis and model reports and other draft references whose contents may change with time. Information that evolves through subsequent revisions of analysis and model reports and other references will be reflected in the LA as the approved analyses of record at the time of LA submittal. Consequently, the Project will not routinely update either this technical basis document or its Key Technical Issue agreement appendices to reflect changes in the supporting references prior to submittal of the LA. May 2004 1-3 No. 10: Unsaturated Zone Transport INTENTIONALLY LEFT BLANK 1-4 No. 10: Unsaturated Zone Transport Revision 1 May 2004 Revision 1 2. RADIONUCLIDES AND TRANSPORT PROCESSES Water from precipitation percolating through Yucca Mountain is the dominant agent determining corrosion of engineered barriers and waste packages, waste dissolution, and radionuclide transport from the repository to the accessible environment. The flow of water, and subsequent transport of radionuclides, is determined by the characteristics of unsaturated zone geology. The amount and rate of radionuclides released from waste packages, and subsequently transported through the unsaturated zone to the water table, are key factors determining the long-term safety of the repository system at Yucca Mountain. Subsurface formations in the unsaturated zone consist of heterogeneous layers of anisotropic, fractured volcanic tuffs with alternating welded and nonwelded ash-flow and ash-fall deposits. The welded units are generally denser, harder, and more intensely fractured than the nonwelded units. The major geologic units of the unsaturated zone beneath Yucca Mountain, from top to bottom, are the volcanic tuff formations of the Paintbrush (Tp) Group, the Calico Hills Formation (Tac), and the Crater Flat (Tc) Group. The lithostratigraphic nomenclature divides the Paintbrush Group into the Tiva Canyon (Tpc), Yucca Mountain (Tpy), Pah Canyon (Tpp), and Topopah Spring (Tpt) tuffs. The Crater Flat Group is divided into the Prow Pass (Tcp), Bullfrog (Tcb), and Tram tuffs (Tct). For purposes of hydrogeologic studies, a separate hydrogeologic nomenclature was developed based on the degree of welding and hydrologic property distributions (CRWMS M&O 2000, Tables 4.7-1 and 8.10-1). The major hydrogeologic units are divided into the Tiva Canyon welded (TCw), the Paintbrush nonwelded (PTn) (consisting primarily of the Yucca Mountain and Pah Canyon members and the interbedded tuffs), the Topopah Spring welded (TSw), the Calico Hills nonwelded (CHn), and the Crater Flat undifferentiated (CFu) units. Based on the hydrogeologic properties of the lithostratigraphic units, Flint (1998) developed a detailed hydrogeologic stratigraphy for use in numerical flow and transport modeling. This correlation between the lithostratigraphic units, hydrogeologic units, and corresponding unsaturated zone flow and transport model layers can be found in Technical Basis Document No. 3: Water Seeping into Drifts, and more details are available in UZ Flow Models and Submodels (BSC 2003b, Section 6.1.1). A summary indicating the correlation between the three different classification systems of the unsaturated zone subsurface is presented in Table 2-1. Note that there is not necessarily one-to-one mapping of the lithostratigraphic units to the flow and transport model layers because geologically dissimilar media can have very similar hydrogeologic behavior. The conceptual model for water flow through the unsaturated zone, developed from extensive field measurements and modeling studies, is shown in Figure 2-1. This model illustrates the geologic control over the flow of water through the unsaturated zone and is the basis for modeling radionuclide transport. Infiltrating water flows through the densely fractured Tiva Canyon welded (TCw) unit mainly through fractures. Within the more porous Paintbrush nonwelded (PTn) unit, most of the water flows through the matrix, where the high storage capacity promotes the dampening of infiltration pulses. Small amounts of flowing water preferentially pass through faults that cut through this unit. The repository lies within the Topopah Spring welded (TSw) unit. Here water flow may become focused into fewer fractures as it approaches the repository horizon, affecting seepage rates into 2-1 May 2004 No. 10: Unsaturated Zone Transport the repository drift and the number of waste packages that are contacted by water. The transport time of radionuclides to the water table will be determined largely by the flow paths (indicated in the figure), the relative dominance of fracture and matrix flow, and the mineralogy of the units. Table 2-1. Major Hydrogeologic Units, Lithostratigraphy, Unsaturated Zone Model Layer, and Hydrogeologic Unit Correlation Major Hydrogeologic Unitsa Tiva Canyon welded (TCw) Paintbrush nonwelded (PTn) Topopah Spring welded (TSw) Lithostratigraphic Nomenclatureb Tpcr Tpcp TpcLD Tpcpv3 Tpcpv2 Tpcpv1 Tpbt4 Tpy (Yucca) Tpbt3 Tpp (Pah) Tpbt2 Tptrv3 Tptrv2 Tptrv1 Tptrn Tptrl, Tptf Tptpul, RHHtop Tptpmn Tptpll Tptpln Tptpv3 Tptpv2 Hydrogeologic Unitc CCR, CUC CUL, CW CMW CNW BT4 TPY BT3 TPP BT2 TC TR TUL TMN TLL TM2 (upper 2/3 of Tptpln) TM1 (lower 1/3 of Tptpln) PV3 PV2 Unsaturated Zone Flow Model Layerd tcw11 tcw12 tcw13 ptn21 ptn22 ptn23 ptn24 ptn25 ptn26 tsw31 tsw32 tsw33 tsw34 tsw35 tsw36 tsw37 tsw38 tsw39 (vitric, zeolitic) No. 10: Unsaturated Zone Transport 2-2 Revision 1 May 2004 Table 2-1. Major Hydrogeologic Units, Lithostratigraphy, and Unsaturated Zone Model Layer, and Hydrogeologic Unit Correlation (Continued) Lithostratigraphic Nomenclature b Tacbt (Calicobt) Tcpuv (Prowuv) Tcpuc (Prowuc) Tcpmd (Prowmd) Tcplc (Prowlc) Tcplv (Prowlv) Tcpbt (Prowbt) Tcbuv (Bullfroguv) Tcbuc (Bullfroguc) Tcbmd (Bullfrogmd) Tcblc (Bullfroglc) Tcblv (Bullfroglv) Tcbbt (Bullfrogbt) Tctuv (Tramuv) Tctuc (Tramuc) Tctmd (Trammd) Tctlc (Tramlc) Tctlv (Tramlv) Tctbt (Trambt) and below Tptpv1 Tpbt1 Tac (Calico) Major Hydrogeologic Unitsa Calico Hills nonwelded (CHn) Crater Flat undifferentiated (CFu) Source: Simmons 2004, Table 7-1. NOTE: a Modified from Montazer and Wilson 1984, Table 1. b BSC 2002. c Flint 1998. d BSC 2003c. No. 10: Unsaturated Zone Transport 2-3 Hydrogeologic Unit c BT1 or BT1a (altered) CHV (vitric) or CHZ (zeolitic) BT PP4 (zeolitic) PP3 (devitrified) PP2 (devitrified) PP1 (zeolitic) BF3 (welded) BF2 (nonwelded) Not Available Not Available Revision 1 Unsaturated Zone Flow Model Layer d ch1 (vitric, zeolitic) ch2 (vitric, zeolitic) ch3 (vitric, zeolitic) ch4 (vitric, zeolitic) ch5 (vitric, zeolitic) ch6 (vitric, zeolitic) pp4 pp3 pp2 pp1 bf3 bf2 tr3 tr2 May 2004 Revision 1 NOTE: Lithostratigraphic nomenclature, hydrogeologic units, and unsaturated zone flow model layers are listed in Table 2-1. Figure 2-1. Overall Water Flow Behavior in the Unsaturated Zone, Including the Relative Importance of Fracture and Matrix Flow Components in the Major Hydrogeologic Units The main hydrogeologic units below the repository are the Calico Hills nonwelded (CHn) and Crater Flat undifferentiated (CFu) units. Both of these units have vitric and zeolitic components that differ in their degree of hydrothermal alteration and subsequent hydrologic properties. The major vitric layers in the CHn (CHv) are in the southern half of the area below the repository. These units have relatively high matrix porosity, low fracture density, and permeability that is [JO1]similar in the fractures and in the matrix. Given the small fracture volumes in these units, a large fraction of the flow occurs through the matrix. The zeolitic rocks have low matrix permeability and some fracture permeability; consequently, a relatively small amount of water may flow through the zeolitic units, with most of the water flowing laterally in perched water bodies and then vertically down faults. Section 3.2 contains a more detailed discussion of unsaturated zone flow and its relationship to transport. Figure 2-2 illustrates the processes affecting the fate and transport of radionuclides from the repository horizon. These processes are described in detail later in this section. Both solute and May 2004 2-4 No. 10: Unsaturated Zone Transport Revision 1 colloidal forms of radionuclides are considered. The important transport processes include advection through the fractures and matrix, dispersion and diffusion within the matrix, sorption, and radioactive decay accounting for the different properties of daughter products. This section continues with descriptions of the various forms of radionuclides considered in transport modeling, followed by discussions of the different transport processes. Figure 2-2. Processes Affecting Transport of Radionuclides May 2004 2-5 No. 10: Unsaturated Zone Transport Revision 1 2.1 RADIONUCLIDES RELEASED INTO THE UNSATURATED ZONE Radionuclides considered in the studies of unsaturated zone transport were selected to represent the species known to occur in substantial concentrations in the radioactive wastes (either as parents or as daughter products of decay). These radionuclides cover the range of transport properties (sorption and diffusion) and half-lives of all representative species listed in Section 5 (in Table 5-2). Because of radioactive decay, the transport of both the parent and daughter products must be accounted for (Figure 2-2f). Daughter products with long half--lives are especially important for performance assessment. Since daughter products may have significantly different transport behavior than parent radionuclides, the migration and fate of all the important members of the decay chain must be considered. In addition, radionuclides may be transported through the unsaturated zone in two different forms: radioactive solutes (which are radionuclides dissolved in water) and radioactive colloids. Figure 2-3 depicts the classification of colloids according to their origin and characteristics (BSC 2003d, Section 6.3.1). In Class I colloids, the entire nonaqueous component of the colloidal particle is radioactive. This class includes true or intrinsic colloids (e.g., plutonium(IV), which are generated from a solute when its concentration exceeds its solubility (Saltelli et al. 1984). Class I colloids also include waste-form colloids, which are produced from the nucleation of colloids from waste-form dissolution and from spallation of colloid-sized waste from alteration products. Waste-form colloids are believed to be one of the most significant contributors to radionuclide transport in the unsaturated zone (BSC 2003d, Section 6.3.1). Class II represents natural colloids, called pseudocolloids, onto which actinides are irreversibly sorbed. Pseudocolloids can be inorganic (e.g., clay, iron oxyhydroxides, silica) or organic (e.g., microbes and humic acids) (Ibaraki and Sudicky 1995, p. 2945) and include all colloidal particles that are neither true colloids nor waste-form colloids. For Class II colloids, radionuclide sorption is, by definition, either irreversible or incorporated into the colloidal structure (e.g., by co-precipitation of radionuclides). Only a portion of the colloidal particle (usually a very small one) is radioactive. In this class of colloids, the actinide remains confined onto the colloid and does not exchange mass with its surroundings (i.e., the liquid phase or adjacent colloids). Class III includes radioactive pseudocolloids in which actinides are reversibly sorbed onto the underlying natural colloid. As in Class II colloids, only a small portion of the colloidal particle is radioactive. However, for Class III colloids, the actinide portion is not confined to the colloid but can exchange mass with its surroundings (i.e., the liquid phase or adjacent colloids). 2-6 May 2004 No. 10: Unsaturated Zone Transport Revision 1 Figure 2-3. Colloid Type and Origin The fate of radioactive colloids is a far more complex and less understood process compared to solutes because of the possible effects of daughter product generation on colloids and the resulting transport properties. It is not known if colloidal size changes when the daughter product is ejected (e.g., due to recoil during alpha decay) or under what conditions the daughter will remain a part of the colloid. The size of pseudocolloids is unlikely to change due to radioactive decay. However, it is not known if the daughter products will remain sorbed or if they will be ejected into the liquid phase. The problem is further complicated by colloid generation, flocculation, and instability, which affect daughter distribution and partitioning in a complex manner (BSC 2003e, Section 6.18). May 2004 2-7 No. 10: Unsaturated Zone Transport Revision 1 2.2 DESCRIPTION OF TRANSPORT PROCESSES 2.2.2 Matrix Diffusion The diffusion of radioactive solutes and colloids into the rock matrix is an important retardation mechanism controlling the radionuclide transfer between the fractures and the rock matrix; it is the only significant retardation mechanism for nonsorbing solutes, such as 99Tc (Figure 2-2a). In the process of matrix diffusion, radionuclides move into the matrix, where water flow is slow and where sorption or filtration (Sections 2.2.2 and 2.2.4) are much more likely to occur because of greater contact areas and times between water and rock (Figure 2-2b). Matrix diffusion removes some radionuclides from the flowing fractures, thus slowing radionuclide transport through the fractures. Diffusive flux within the matrix of a given species is a function of its 2.2.1 Advective and Dispersive Transport Water flowing through the unsaturated zone can transport radionuclides from the emplacement location to the water table (Figure 2-1). Fracture flow is expected to be the dominant transport mechanism in most hydrogeologic units except for the vitric CHv unit and perched water zone because of high fracture permeability compared to matrix permeability, limited fracture pore volumes, limited fracture–matrix contact areas, and short contact times between the radionuclide–carrying liquid phase and the rock matrix (Figure 2-2a). Fracture-dominated advective transport is expected to lead to the earliest arrivals at the water table due to fast water flow and because the aforementioned factors can adversely affect the two main retardation mechanisms: (1) matrix diffusion and (2) immobilization by means of solute sorption onto or colloid filtration by the matrix (BSC 2003e, Section 6). While advective flow is predominantly downward in response to gravity some lateral advection is also expected in response to lateral flow diversion at the boundaries of hydrogeologic units that have sharp contrasts in their hydraulic properties (BSC 2003b, Section 6.2.2). Flow diversion may also occur in the perchedwater bodies of the unsaturated zone. Laterally diverted flow ultimately finds a pathway to the water table through other more permeable zones, such as faults or connecting fractures. Mechanical dispersion arises from localized velocity variations and, combined with diffusion, constitutes hydrodynamic dispersion (Figure 2-2d). Longitudinal dispersion (i.e., dispersion in the direction of flow) may result in earlier arrivals of smaller concentrations of radionuclides at the water table ahead of the mean concentration front. However, dispersion is not expected to play a significant role in radionuclide transport in the system dominated by fracture flow. Additionally, dispersion effects are small because matrix diffusion spreads the arrivals of radionuclides more significantly compared to spreading by mechanical dispersion. Finally, the repository emplacement area is very broad (approximately 5,000 m by 2,000 m) relative to the distance to the water table (about 300 m); this configuration has been shown to minimize lateral dispersion effects (BSC 2003b, Section 6.7.1). Based on the results of inverse modeling (BSC 2003f), the calibrated properties indicated that, in the vitric CHv units, matrix flow dominates because of similar permeability of fractures and matrix (Figure 2-1) and the low fracture volumes. This, combined with the high matrix pore volume, results in much slower transport velocities in addition to larger water–rock contact areas and longer radionuclide–matrix contact times (BSC 2003e, Section 6). 2-8 May 2004 No. 10: Unsaturated Zone Transport Revision 1 molecular properties (e.g., electric charge, size, and diffusion coefficient), its concentration gradient, and other properties, such as temperature, the matrix pore structure, and water saturation of the pore space (BSC 2003e, Section 6.2). In general, matrix diffusion is less for colloids than for solutes, because the large size of the colloids: (1) reduces the colloid diffusion coefficient; (2) reduces the number of colloids entering the matrix because of pore size exclusion; and (3) enhances advection by guiding the colloids to the center of the fracture pores where velocities are higher than the average water velocity (Figure 2-2c) (Ibaraki and Sudicky 1995, p. 2948). 2.2.3 Sorption Sorption describes a combination of chemical and physical interactions between dissolved radionuclides and the rock matrix. Sorption removes a portion of the dissolved species from the mobile liquid phase and transfers it to the solid phase. The solid phase includes both the immobile rock matrix and mobile (Figure 2-2b) or immobile (Figure 2-2e) colloids. If the solid phase is immobile, sorption results in larger concentration gradients between the sorption sites and the bulk matrix, thereby enhancing diffusive fluxes between fractures and matrix and reducing the rate of solute transport. The sorption behavior of radionuclides is usually described by the distribution coefficient (Kd), which quantifies the partitioning of radionuclides between the solid and aqueous phases. At the same temperature, repetitive experiments with different concentration of solutes can construct a family of points forming a linear equilibrium isotherm, based on which the Kd can be estimated for the solutes of interest. The Kd values for Yucca Mountain studies are obtained from laboratory batch experiments using crushed tuffs with a particle size of 75 to 500 µm (BSC 2003e, Attachment I) contacted with groundwater (or simulated groundwater) representative of the site, spiked with one or more of the elements of interest. Sorption experiments have been carried out as a function of time, element concentration, atmospheric composition, particle size, and temperature. In some cases, the solids remaining from sorption experiments were contacted with unspiked groundwater in desorption experiments. The sorption and desorption experiments together provide information on the equilibration rates of the forward and backward sorption reactions (see Section 4.1). The primary controls on sorption behavior in the unsaturated zone flow system include the nature of mineral surfaces in the rock units contacted by the flowing water, the pore-water and perched water chemistry, the specific sorption mechanisms for each radionuclide, the concentrations of the various radionuclides in water, and their speciation and redox state. Effective Kd values obtained from batch experiments involving high concentration solutions will tend to underestimate the field Kd values if the expected field concentrations are low and nonlinear or kinetic sorption prevails (BSC 2003e, Section 6.1.2.3). Conversely, batch experiments using crushed rock samples may overestimate the Kd values (compared to intact rock samples and for the same solute concentration) if kinetic effects are involved and the contact time is short. Even though procedures were used to minimize the inclusion of fines, crushing of the material creates new contact surfaces and increases the radionuclide accessibility to pores that may not contribute to sorption on intact rocks. The difference in the Kd estimates May 2004 2-9 No. 10: Unsaturated Zone Transport Revision 1 from the two sample types vanishes for sufficiently long contact times. The deterministic approach in the Yucca Mountain studies tends to underestimate the Kd values because it involves an averaging of Kd values obtained from experiments using both low and high solution concentrations, while the release concentrations from the repository are expected to be low. The underestimation of Kd values produces an earlier arrival prediction of radionuclide mass to the water table. 2.2.4 Colloid Transport Colloid filtration is categorized into two processes: (1) physical-chemical deposition and (2) straining. Colloid deposition onto rock surfaces as water flows through its pores (Figure 2-2b) is commonly assumed to occur in two steps: (1) transport of colloids to matrix surfaces by Brownian diffusion, interception, or gravitational sedimentation (i.e., colloid-matrix collision); and (2) attachment of colloids to matrix surfaces. The fraction of collisions resulting in attachment is called the attachment efficiency and is strongly influenced by interparticle forces between colloids and matrix surfaces, such as van der Waals and electric double-layer interactions, steric stabilization, and hydrodynamic forces (Kretzschmar et al. 1995, p. 435). Kretzschmar et al. (1997, p. 1129) demonstrated that colloid deposition generally follows a first-order kinetic rate law, and they experimentally determined the corresponding collision efficiencies. Once attached, colloid detachment is generally slow, and attachment can appear to be irreversible. Colloid straining (Figure 2-2c) can also affect the distribution and transport of colloids. Straining mechanisms can be classified according to the relative size of the colloid as conventional straining (if the colloid is larger than the pore throat diameter or the fracture aperture, pore size exclusion accounted for in the study of colloid transport in Section 6), and film straining (if the colloid is larger than the thickness of the adsorbed water film coating the grains of the rock) (Wan and Tokunaga 1997, p. 2413). McGraw and Kaplan (1997, p. 5.2) found a very strong dependence of filtration on the colloid size under unsaturated conditions. Colloid removal increased exponentially with colloid size, and the decrease in colloid mobility at low volumetric water contents was attributed to resistance due to friction (as the colloids were dragged along the sand grains). Wan and Tokunaga (1997, pp. 2413 and 2419) found that if the water saturation of the pore space is lower than a critical saturation value, colloids can only move in the thin film of water that lines the grain boundaries. Colloid retardation increased as the ratio of the water film thickness to the colloid diameter decreased. Two classes of colloids, Class I and Class II (Section 2.1), were considered in the transport study (BSC 2003e, Section 6.18). Class I includes true colloids and waste form colloids, in which the entire non-aqueous component of the colloid is composed of the radioactive substance. Those were taken to have the properties of PuO2 and are subject to radioactive decay. Note that the assumption of PuO2 colloids indicates that radioactivity follows the decay of 239Pu. This does not affect their overall transport behavior (a strong function of size) or limit the validity of observations. Additionally, the Class I colloid size and density were considered invariable during the simulation. This stipulation means that matrix diffusion does not increase in response to a smaller colloid diameter and colloid filtration (the parameters of which are based on the actinide mass per unit volume) does not decrease due to lower concentrations. In both colloidal classes considered in this study, four different colloid sizes were considered (see Section 6.2). May 2004 2-10 No. 10: Unsaturated Zone Transport Revision 1 3.1 CLIMATE AND INFILTRATION Climatic conditions play a critical role in transport. They are key for estimating the net infiltration and deep percolation in the unsaturated zone, and ultimately in estimating groundwater flux. Under a wetter climatic regime, the increased precipitation leads to increased 3. RELATIONSHIP OF RADIONUCLIDE TRANSPORT TO OTHER COMPONENTS OF THE REPOSITORY SYSTEM Unsaturated zone transport processes are impacted by various components of the repository system. Precipitation and near-surface processes determine the infiltration of water to the unsaturated zone (Figure 2-1). The water that percolates below the root zone (i.e., net infiltration) follows pathways that are (among other factors) determined by the geology of Yucca Mountain. As infiltration approaches the repository horizon, flow is focused into fewer fractures. Even though there is a lack of field evidence to show the flow focusing, 36Cl data (Fabryka-Martin et al. 1996) and field observation (Wang et al. 1998) show that there are discrete flow features through fractures. The percolation flux distribution affects the fraction of waste emplacement drifts encountering seepage, the seepage rate, and thus potential water–waste contact. Flow through the engineered barrier system within the repository impacts water–waste contact and provides the source term for radionuclide transport. Flow diversion around the drifts creates a shadow zone below the repository, which determines the transfer mechanism and release rates of radionuclides from the waste emplacement drifts to the natural system, and thus affects the nature of the radionuclide source term for transport through the unsaturated zone. Figure 2-1 also shows other hydrologic features that impact radionuclide transport below the repository, including perched water and the distinctly different flow patterns through the northern and southern portions of the Calico Hills unit. The unsaturated zone flow model and its submodels are developed to simulate past, present, and future hydrogeologic, geothermal, and geochemical conditions and processes within the Yucca Mountain unsaturated zone to support various TSPA-LA activities. A detailed analysis of simulated percolation fluxes at the repository level and at the water table was conducted for nominal simulation scenarios of unsaturated zone flow fields (BSC 2003b). Flow modeling for the unsaturated zone has indicated whether fracture or matrix flow dominates in a given hydrogeologic unit. With this understanding, the relative importance of various radionuclide transport processes and the potential of the unsaturated zone for retarding the transport of radionuclides to the water table can be evaluated. Finally, the unsaturated zone transport predictions provide the source terms for saturated zone predictions of transport to the accessible environment. The repository-system components that relate to radionuclide transport are described in Figure 1-1. These components include climate and infiltration, unsaturated zone flow, the engineered barrier system, and saturated zone transport and are summarized in the following sections. Technical basis documents for each component provide additional detailed information. The effects of thermal loading and the engineered barrier system in the vicinity of the repository are discussed in the current document. The interaction among repository-system components with radionuclide transport processes is a key part of synthesizing information for the development of transport models and is discussed in more detail in Section 5. May 2004 3-1 No. 10: Unsaturated Zone Transport Revision 1 water percolation through the unsaturated zone and the repository, resulting in faster water flow and earlier radionuclide arrivals at the water table. With faster flow, transport is accelerated by the decreased travel and contact times that limit radionuclide retardation by the unsaturated zone. Forecasting of climatic conditions at Yucca Mountain indicates that the present-day climate persists for 400 to 600 years, followed by a warmer and much wetter monsoon climate for 900 to 1,400 years (USGS 2001, Table 2, p. 67). This monsoon climate is followed by a cooler and wetter glacial transition climate for the remaining 8,000 to 8,700 years (USGS 2001, Table 2, p. 67). Maps of steady-state net infiltration for the present-day, monsoon, and glacial transition climate states over the Yucca Mountain region were prepared from numerical simulations (Figure 3-1). These include lower-bound, mean, and upper-bound conditions for each of these climatic states. The distributions of infiltration indicated by the maps were then used to calculate space- and time-averaged steady-state infiltration rates for each of the climate scenarios (USGS 2003, Section 6.4 and 6.11). The concept behind numerical simulations of net infiltration is based on the solution of the water balance, plug-flow equation using the INFIL V2.0 numerical code. For each of the watersheds, the INFIL code performs daily simulations of net infiltration over all model cells, using the calculations of the evapotranspiration rate and surface runoff/run-on. The main types of data used for numerical modeling of infiltration rate include the precipitation and temperature records, the water storage capacity of the root zone, the field capacity of soils, and hydraulic conductivity of the soil and bedrock. Results of field and modeling investigations over the Yucca Mountain region (USGS 2003, Section 6.11) show that the net infiltration rate varies greatly in space and time depending on storm amplitudes, duration, and frequencies. In very wet years, infiltration pulses to the unsaturated zone of Yucca Mountain may occur over a relatively short time period (Bodvarsson et al. 1999, p. 10). Where the soil thickness decreases and bedrock crops out, net infiltration increases because fast preferential flow through rock fractures exceeds the depth of evapotranspiration (Flint and Flint 1995, p. 14). Spatial and temporal variability in net infiltration at Yucca Mountain are caused by episodic storm and precipitation events (Hevesi et al. 1994) as well as the heterogeneous nature of the topsoil layer and topography. Near-surface infiltration data (USGS 2003, Section 6.11) suggest that significant infiltration occurs only every few years. In very wet years, infiltration increases to hundreds of millimeters per year during a relatively short time. Climatic conditions and infiltration at Yucca Mountain and their impact on the repository system behavior are discussed in detail in Technical Basis Document No. 1: Climate and Infiltration. May 2004 3-2 No. 10: Unsaturated Zone Transport Source: DTN: GS000308311221.005. NOTE: The TSPA-LA model grid is presented in Section 5, Figure 5-1. Figure 3-1. Plan View of Net Infiltration Distributed over the Three-Dimensional Unsaturated Zone TSPA-LA Model Grid for the Present-Day (Base-Case) Mean Infiltration Scenario 3-3 No. 10: Unsaturated Zone Transport Revision 1 May 2004 Revision 1 3.2 UNSATURATED ZONE FLOW Accurate determination of the relative importance of fracture and matrix flow components is especially important for chemical transport processes. Flow in fractures is typically much faster than flow in the matrix, leading to much shorter transport times for radionuclides and other chemicals in fractures compared to the matrix. The characteristic flow behavior in each of the major hydrogeologic units shown in Figure 2-1 is supported by extensive field observations or modeling studies. In general, flow is primarily fracture-dominated in the densely welded and intensely fractured TCw and TSw hydrogeologic units and matrix-dominated in the nonwelded, less intensely fractured PTn unit, where flow is damped. This conceptual understanding was the basis for the development of the site-scale unsaturated zone flow model. The high density of interconnected fractures and low matrix permeabilities in the TCw unit (BSC 2003g, Sections 6.1 and 6.2) are considered to give rise to significant water flow in fractures and limited matrix imbibition (water flow from fractures to the matrix). Thus, episodic infiltration pulses are expected to move rapidly through fracture networks, with little attenuation by the matrix. The relatively high matrix permeabilities and porosities and low fracture densities of the PTn unit (BSC 2003g, Sections 6.1 and 6.2) should convert the predominant fracture flow in the TCw to dominant matrix flow within the PTn (BSC 2003b, Section 6.2.2). The dominance of matrix flow and the relatively large storage capacity of the matrix (resulting from its high porosity and low saturation) give the PTn significant capacity to attenuate infiltration pulses. Faults and geological structures may cut through the entire PTn unit at some locations, leading to fast flow paths when the localized tuff matrix is not dry enough to imbibe water flowing in the fractures. However, recent modeling studies support the existence of lateral flow within the PTn (BSC 2003b, Section 6.6.3). These modeling results show that this lateral flow pattern through the PTn has a large impact on percolation flux distribution in the repository horizon. These percolation fluxes and their distributions at the repository level indicated that there exists a certain amount of large-scale lateral flow or diversion by the PTn unit. The repository resides within the TSw hydrogeologic unit. The spatially and temporally damped fluxes from the PTn flow into the fractures of the TSw, where the flow becomes focused into fewer fractures or faults as it approaches the repository horizon because of lower fracture density (BSC 2003g, Section 6.1.2). Flow focusing and flow channelization in the fractures may affect seepage rates into repository drifts and thus the number of waste packages potentially contacted by water. Field evidence at Yucca Mountain suggests that preferential flow pathways may exist in the unsaturated rocks of Yucca Mountain. Preferential flow pathways are in part supported by elevated levels of 36Cl, attributed to atmospheric nuclear tests conducted in the 1950s and 1960s, which have been reported at several locations in the Exploratory Studies Facility (ESF) (Fabryka-Martin et al. 1996, Section 9). Focusing flow along these preferential paths or wellconnected fracture networks plays an important role in controlling patterns of percolation through highly fractured tuffs, such as the TSw unit, and have a direct impact on seepage into drifts (BSC 2001a, Section 6.4.2). Flow behavior below the repository is especially important for predicting radionuclide transport from the repository horizon to the water table because transport paths follow the pattern of water flow. The main hydrogeologic units below the repository are the CHn and CFu units. Both of May 2004 3-4 No. 10: Unsaturated Zone Transport Revision 1 these units have vitric and zeolitic components that differ in their degree of hydrothermal alteration and subsequent hydrological properties. The zeolitic rocks have low matrix permeability and some fracture permeability; consequently, a relatively small amount of water may flow through the zeolitic units, with most of the water flowing laterally in perched water bodies and then vertically down faults (BSC 2003b, Section 6.6.3). Perched water zones have been reported in a number of boreholes within the lower portion of the TSw unit and within the upper portion of the CHn unit (BSC 2003b, Section 6.2.2). These perched-water bodies were found primarily in the northern part of the repository area, where lower-permeability, sparsely fractured zeolitic rock units predominate. The occurrence of perched water suggests that certain layers of the lower TSw (e.g., the basal vitrophyre) and the upper zeolitic CHn serve as barriers to vertical flow. On the other hand, the vitric units have relatively high matrix porosity, and a matrix permeability that is similar to their fracture permeability. Because of (1) the permeability parity in the matrix and in the fractures and (2) the limited volume of the fractures (compared to the matrix), simulation results indicate that matrix flow dominates in these vitric units (BSC 2003h). Test results within the CHn at the Busted Butte underground facility show that water flow and radionuclide transport occur mainly within the matrix of the CHn, implying that fracture flow is limited in this unit (BSC 2003h). The extent of fracture–matrix interaction is a key factor in determining flow and transport processes in the unsaturated zone. Fracture–matrix interaction refers to flow and transport (or mass exchange) between fractures and the matrix. Because of their different hydrologic properties, distinct flow and transport behavior occurs in each component. Modeling results and field observations show limited fracture–matrix interaction in welded units at Yucca Mountain (BSC 2003b, Section 6.6.3). Fingering flow in fractures in unsaturated fractured rocks is believed to be a major reason for limiting fracture–matrix interaction. To incorporate the effects of fingering flow into modeling of flow and transport in unsaturated fractured rocks, the active fracture model was developed by Liu et al. (1998, pp. 2633 to 2646), as documented in Analyses of Hydrologic Properties Data (BSC 2003g, Section 6.6). In the model, only a fraction of connected fractures are considered to conduct liquid water as a result of fingering flow at a fracture network scale and within individual fractures. Major faults may have the potential to significantly affect the flow processes at Yucca Mountain. A fault is considered to serve as a localized, fast-flow conduit for liquid water, especially below the repository. Low-permeability layers (or perched water zones) at the base of the TSw and in the CHn may laterally divert a considerable amount of flow to major faults, which may focus flow downward to the water table. However, it is also possible that alteration within or along faults in the CHn and CFu reduces their permeability, increasing water travel times from the TSw to the water table. Nevertheless, to be conservative, faults below the repository have been treated as localized fast-flow paths in the current unsaturated zone flow model (BSC 2003b, Section 6.1.1). The site-scale unsaturated zone flow model (BSC 2003b) has been developed, calibrated, and verified based on the geology of the site and field observations. The unsaturated zone flow model is a three-dimensional dual-permeability model. A variety of data (including matrix saturation and water potential data, pneumatic data, perched-water data, temperature data, and geochemical data) has been used for calibrating and confirming the model. Confidence in the May 2004 3-5 No. 10: Unsaturated Zone Transport Revision 1 model is increased through comparison against a substantial amount of field data not used for model calibration. These confidence building activities include checking for consistency between modeling results with hydrologic data, geochemical data, and data collected from in situ tests. Effects of thermal processes and the resultant coupled processes on unsaturated zone flow are also treated. The site-scale unsaturated zone flow model generates unsaturated zone flow fields used directly in TSPA analyses. Unsaturated zone flow and its relevance for the repository system behavior are discussed in detail in Technical Basis Document No. 2: Unsaturated Zone Flow. 3.3 SHADOW ZONE Diversion of percolating water around a waste emplacement drift results in an environment of greatly diminished flow inside the drift compared to the adjacent undisturbed formation. In such an environment, diffusion is practically the only radionuclide transport mechanism to the surrounding rock. Because of the relatively large water content of the matrix in comparison with the fractures, diffusive releases from waste emplacement drifts are strongly partitioned to the matrix. Under these conditions, transport times to the water table increase by orders of magnitude compared to fracture release (BSC 2003i, Section 6.3). Simulation results show that transport is dominated by matrix advection and that most of the radionuclide mass is retained within the matrix at breakthrough at the water table. Consequently, the effects of the drift shadow on transport extend into the saturated zone. The effects of variations in the flow field and fracture–matrix interaction in the vicinity of a waste emplacement drift, and sensitivity to the corresponding parameters, were investigated by (BSC 2003i). These simulations indicated that, when radionuclides were released into the rock matrix, transport was not significantly affected when flow beneath the emplacement drift was substantially reduced. Sensitivity calculations have shown that the main effect of the drift shadow on transport is that radionuclide transport is initiated in the matrix (BSC 2003i, Section 6.3). This leads to slow transport through the unsaturated zone because transport is largely confined to the lowpermeability matrix. Contrary to the effects on flow, the impact of the drift shadow on transport is not significant when the radionuclides are released into the matrix. Varying degrees of fracture–matrix interaction in the dual-permeability model used in these simulations (BSC 2003i) are shown to have a significant influence on transport in the drift shadow. 3.4 SATURATED ZONE TRANSPORT The radionuclides that enter the saturated zone are expected to be retarded over a spatial and temporal scale that is dependent on the degradation modes and rates of the engineered barriers. For example, it is possible that the engineered barriers fail either over a broad temporal scale (ranging from thousands to hundreds of thousands of years) due to natural degradation processes, or over a relatively short time interval associated with a low-probability disruptive event, such as a large seismic event or a volcanic event. The spatial scale over which radionuclides enter the saturated zone may be: (1) relatively confined to an area near vertically below each degraded waste package sits (for cases where flow is predominantly vertical through the unsaturated zone); May 2004 3-6 No. 10: Unsaturated Zone Transport Revision 1 (2) concentrated at locations where the bulk of the unsaturated groundwater flow intersects the water table; or (3) dispersed over a significant fraction of the total repository footprint of several square kilometers. Uncertainty in the timing and spatial extent of the radionuclides that enter the saturated zone is addressed by considering a range of locations and combining the advective-dispersive transport times within the saturated zone to those times when radionuclides are predicted to reach the saturated zone. This abstraction process is described in SZ Flow and Transport Model Abstraction (BSC 2003j). Saturated zone flow and transport processes are represented by different conceptual and numerical models, including models of groundwater flow at the regional and site scale and models of radionuclide transport. The bases of these models are derived from site-specific in situ observations, as well as field and laboratory tests to determine the relevant parameter values used in these models. The most likely pathway for radionuclides to reach the accessible environment is through the uppermost groundwater aquifers below the repository. These aquifers, collectively referred to as the saturated zone, delay the transport of radionuclides to the accessible environment and reduce the concentration of radionuclides before they reach the accessible environment. Delay in the release of radionuclides to the accessible environment allows radioactive decay to diminish the mass of radionuclides that are ultimately released. Dilution of radionuclide concentrations in the groundwater used by the potential receptor occurs during transport and in the process of extracting groundwater from wells. The key processes that affect the performance of this barrier include advection, sorption, diffusion (in particular matrix diffusion), hydrodynamic dispersion, decay and ingrowth, and filtration of colloids carrying radionuclides. Analyses conducted using the saturated zone transport model indicate that the saturated zone is expected to provide significant retardation to the transport of radionuclides to the accessible environment within the 10,000-year period of regulatory concern for the repository at Yucca Mountain. The expected behavior of the saturated zone system is to delay the transport of sorbing radionuclides and radionuclides associated with colloids for many thousands of years, even under wetter climatic conditions in the future. Nonsorbing radionuclides are expected to be delayed for hundreds of years during transport in the saturated zone. Flow and transport processes in the saturated zone are discussed in detail in Technical Basis Document No. 11: Saturated Zone Flow and Transport. 3.5 TEMPERATURE EFFECT ON TRANSPORT The effects of higher temperature on transport are described by its potential impact on flow (the vehicle of advective transport), and on the magnitude of the main transport parameters. The discussion in the ensuing section indicates that large-scale thermally induced alteration of the hydrologic properties (and, consequently, of large-scale advective transport) of the unsaturated zone is expected to be minimal. However, higher temperatures in the immediate vicinity of the repository can have a drying effect that significantly affects transport by distorting the flow field, enlarging the shadow zone, and leading to reduced releases. Temperature has an effect on the magnitude of the transport parameters. 3-7 May 2004 No. 10: Unsaturated Zone Transport Revision 1 3.5.1 Effect on Large-Scale Flow and Advective Transport Because radionuclide transport is controlled by water flow through the unsaturated zone, the thermal effects on flow will affect transport. The mountain-scale thermal-hydrologic (TH), thermal-hydrologic-chemical (THC), and thermal-hydrologic-mechanical (THM) process models were developed to assess mountain-scale hydrologic, chemical, and mechanical changes and to predict unsaturated zone flow behavior in response to heat release by radioactive decay of radioactive waste. The three-dimensional TH model predicts that in a ventilated repository, heating will have, in general, only a limited impact on far-field flow fields, with thermally enhanced flux zones not extending 30 m beyond the repository. Without ventilation, on the other hand, thermally impacted flux zones extend as far as 100 m from the repository. Mineral dissolution and precipitation leading to changes in hydrologic properties may affect the percolation flux to the repository horizon, impacting seepage into and flow below the repository, which will impact radionuclide transport in the unsaturated zone. The results of the THC simulations (BSC 2003k, Section 6.4.4) indicate that mineral precipitation/dissolution will not significantly affect the hydrologic properties and the percolation flux compared to the effects caused by TH processes alone. Changes to water chemistry, mineralogy, and hydrologic properties in the ambient temperature regions are minimal over the 10,000 years of the simulation. The mountain-scale THM model is capable of assessing the magnitude and distribution of changes in hydrologic properties and of analyzing the impact of such changes on the large-scale vertical percolation flux through the repository horizon (BSC 2003l, Section 6.5). The result shows that maximum THM-induced changes in hydrologic properties occur at around 1,000 years after waste emplacement, when average temperature in the mountain reaches maximum values. THM-induced changes in large-scale hydrologic properties have no significant impact on the vertical percolation flux through the repository horizon. The implication of these results is that if radionuclide releases begin 2,000 years or later after the waste package emplacement in the repository, temperature effects on flow and advective transport will be negligible. 3.5.2 Effects on Transport in the Shadow Zone In the shadow zone below the drift (Figure 3-2a), the transport of radioactive solutes is retarded because of the reduced water saturations (BSC 2003m, Section 6.7). The heat generated by the decaying radioactive waste can significantly accentuate this condition by further reducing water saturations and possibly leading to dryout regions in the vicinity of the waste emplacement drifts. Low water saturations reduce water velocities in the shadow zone (into which water moves mainly in response to capillary pressure differentials) and significantly lowers diffusive fluxes across the fracture–matrix interfaces. The effect of the shadow zone on the transport of colloids is expected to be far more significant, because advective fluxes are low, and diffusive fluxes (the only mechanism capable of transporting radionuclides past the boundaries of the shadow zone) are depressed, because the colloid diffusion coefficient is generally orders of magnitude lower than the molecular diffusion coefficient of solutes. 3-8 May 2004 No. 10: Unsaturated Zone Transport Revision 1 3.5.3 Effects on the Magnitude of the Transport Parameters The transport parameters and processes that can be affected by a temperature increase are: (1) the diffusion coefficient D0 of the dissolved or colloidal species; (2) the sorption parameters of the dissolved species or the filtration parameters of the suspended colloid; (3) the near-field hydrology and the extent and behavior of the shadow zone; and (4) the ambient hydrologic and chemical conditions, which can contribute to thermal alteration of the rock near the repository. These temperature-related consequences are depicted in Figure 3-2. Figure 3-2. Effects of Heat and the Engineered Barrier System on Transport Processes for Solutes and Colloids in the Shadow Zone May 2004 3.5.3.1 Temperature Effects on Diffusion Ambient temperature at the repository horizon at Yucca Mountain is about 25°C, and the maximum temperature at which boiling of the fracture and pore fluids will occur ranges from 95°C to 110°C. Temperature enhances diffusion by increasing Brownian motion (Figure 3-2c). This effect is quantified by an increase in the D0 of radionuclides. Based on a relationship proposed by Robin et al. (1987, pp. 1105 to 1106), an increase in temperature from 20°C (at the top of the unsaturated zone domain) to 30°C (at the water table) leads to an increase of D0 by about 30%. The effect is stronger when considering the temperature changes induced by wastegenerated heat. 3-9 No. 10: Unsaturated Zone Transport Revision 1 3.5.3.2 Temperature Effects on Sorption and Filtration The effect of temperature on sorption is less clear. Temperature increases generally lead to an increased Kd, the value of which depends on the species involved, the nature of the sorbing substrate, the composition of the aqueous phase, and the assumed sorption mechanism (Figure 3-2c). The critical parameter defining the temperature dependence of Kd is the enthalpy of sorption, .Hr, which is usually relatively small in magnitude. Changes in the degree of sorption between ambient temperature at the repository horizon at Yucca Mountain, about 25¡ÆC, and the maximum temperature at which boiling of the fracture and pore fluids will occur, 95¡ÆC to 110¡ÆC, will be relatively small (by about a factor of 10, depending on the radionuclide species under consideration), when compared to the estimated ranges of sorption coefficients (see Table 5-1). For strontium, europium, and uranium(VI), Kd increases 1.6 to 3.8 times due to temperature increases from 25¡ÆC to 95¡ÆC (BSC 2003n). When modeling solute radionuclide transport in the near field, it can be assumed within experimental error that the Kd values of all cited radionuclides are not temperature dependent. This simplifying assumption can be considered conservative because, for most species, the mean value of .Hr is small and likely within the limits of uncertainty imposed by the uncertainties of other model parameters. Laboratory studies (BSC 2003n) confirmed the validity of this approach over a wide range of radionuclides. Colloid filtration (deposition) generally follows a kinetic process (BSC 2003e, Sections 6.2.3 and 6.16.2). An increase in temperature increases the forward filtration coefficient, ¥ê+, indicating an increase in the filtration (i.e., deposition) rate (BSC 2003e, Equation 6-31). There is no information on the effect of temperature on the backward filtration coefficient, ¥ê. . 3.5.3.3 Temperature Effects on the Sorptive Properties of the Host Rock Higher temperatures may thermally alter the devitrified and zeolitic layers below the repository horizon, affecting the sorption of radionuclides (Figure 3-2b). Significant changes in the sorptive properties of the host rocks and zeolites are predicted to occur under the thermal, hydrologic, and geochemical conditions prevailing during the thermal perturbation caused by waste emplacement (BSC 2003o). In the Calico Hills Formation, the dominant zeolitic zone underlying the repository horizon, the temperature is expected to rise to about 70¡ÆC and about 15 vol % of the clinoptilolite is predicted to alter to more stable secondary feldspars and stellerite. Partial dehydration of the clinoptilolite may occur (Carey and Bish 1996; 1997). However, the expected temperature excursion would be insufficient to alter coexisting opal-CT to quartz. Opal-CT maintains a high silica activity in the system and curtails alteration of the zeolites to more stable, but less sorptive phases (Duffy 1993a; 1993b). Despite current modeling uncertainties relating to the persistence of zeolites, the determination that sorption coefficients of given radionuclides have temperature dependencies, as discussed above, remains a valid basis for predicting radionuclide behavior at Yucca Mountain. It has been determined that coupled THC processes will be inconsequential because the duration and magnitude of the thermal pulse and kinetic factors affecting the metastabilities of extant phases are too small to cause significant alteration and the water fluxes are not sufficiently large to dissolve and redeposit significant quantities of the minerals (BSC 2003k, Section 7.3). The 3-10 May 2004 No. 10: Unsaturated Zone Transport Revision 1 inherent stability of zeolites under the geochemical and hydrologic conditions prevailing during the thermal perturbation due to waste emplacement is such that no significant changes in the sorptive properties of the host rocks and zeolites are expected during this period. Thus, it can be concluded that there is no need to account for zeolitic alterations in the simulations. For the radionuclides with temperature dependent Kd values, a simple modification of the sorption coefficient would be sufficient to account for the radionuclide sorptive behavior of the rock matrix at Yucca Mountain. May 2004 3-11 No. 10: Unsaturated Zone Transport INTENTIONALLY LEFT BLANK 3-12 No. 10: Unsaturated Zone Transport Revision 1 May 2004 Revision 1 4. LABORATORY MEASUREMENTS AND FIELD OBSERVATIONS 4.1 LABORATORY SORPTION MEASUREMENTS Measurements of the sorption distribution coefficient, Kd, in the unsaturated zone rocks are critical in the determination of sorption-induced retardation of radionuclide transport. All modeling predictions of radionuclide fate and transport through the unsaturated zone are based on these measurements. This section provides the bases for the derivation of sorption coefficients and their probable ranges used in the unsaturated zone transport calculations (BSC 2003e). The sorption coefficient data on which the distributions are based are obtained in laboratory experiments in which crushed rock samples from the Yucca Mountain site are contacted with groundwater (or simulated groundwater) representative of the site and spiked with one or more of the elements of interest. Sorption experiments have been carried out as a function of time, element concentration, atmospheric composition, particle size, and temperature. In some cases, the solids remaining from sorption experiments were contacted with unspiked groundwater in desorption experiments. The sorption and desorption experiments together provide information on the equilibration rates of the forward and backward sorption reactions. For elements that sorb primarily through surface complexation reactions, the experimental data are augmented with the results of calculations using PHREEQC V2.3 (BSC 2001b). The inputs for the calculations include groundwater compositions, surface areas, binding constants for the elements of interest, and thermodynamic data for solution species. These calculations provide a basis for interpolation and extrapolation of the experimentally derived sorption coefficient data set. The primary controls on sorption behavior of the elements of interest in the unsaturated zone flow system include the detailed characteristics of mineral surfaces in the rock units through which water flows from the repository to the saturated zone. It also includes the detailed chemistry of pore waters and perched waters in the unsaturated zone along flow paths, the sorption behavior of each element, and their concentrations in the groundwater. These parameters are discussed in the following sections. There are three dominant rock types in the unsaturated zone along potential flow paths from the repository to the saturated zone: devitrified tuff, zeolitic tuff, and vitric tuff (BSC 2003c, Section 6). Devitrified tuff is composed primarily of silica (quartz and cristobalite) and alkali feldspar. It may also contain trace amounts of mica, hematite, calcite, tridymite, kaolinite, and hornblende, as well as minor amounts (less than 25%) of smectite and/or zeolite. For the purpose of this analysis, sorption coefficient distributions for devitrified tuff are based on data obtained from samples that are composed primarily of silica phases and feldspar with only trace amounts of other phases. Sorption coefficient distributions for zeolitic tuff are taken exclusively from samples that contain more than 50% zeolite, with the balance made up of clay, silica phases, alkali feldspar, and/or glass. Zeolitic tuffs have significantly higher surface areas compared to other tuffs (Triay et al. 1996, p. 62). For vitric tuffs, sorption coefficient distributions are based on samples that contain more than 50% glass, with the remainder composed predominantly of feldspar, silica phases, zeolites, and clay. 4-1 May 2004 No. 10: Unsaturated Zone Transport Revision 1 Sorption coefficients for the radionuclides of interest can be a function of the concentrations of the radionuclides present in solution. In most cases, experiments were carried out over a range of concentrations up to the solubility limit. Above the solubility limit, a solid phase incorporating the element of interest is precipitated out of solution. Therefore, the concentration of an element in solution cannot rise much higher than the solubility limit. Because experiments have been carried out at concentrations up to the solubility limit for most elements of interest, the experimental results and the probability distributions derived from them include this dependency. The only element for which the experimental concentrations did not approach the solubility limit was cesium, which has a very high solubility in Yucca Mountain waters (BSC 2003p, Section 6.17); cesium concentrations in Yucca Mountain groundwaters will not have a solubility limitation (BSC 2003p, p. I-27) In the unsaturated zone, two rather distinct water types exist in the ambient system. One is perched water and the other is pore water. Perched water is generally more dilute than pore water. The J-13 (water from the main supply well for activities) and UE-25 p#1 (groundwater obtained from Paleozoic carbonate rocks at depth beneath Yucca Mountain) waters were used in sorption experiments as end-member compositions intended to bracket the impact of water composition on sorption coefficients. The two selected waters bracket the chemical composition of pore water and perched water found in the unsaturated zone. Because water from the UE-25 p#1 well was not available to the experimental program at all times, synthetic p#1 water was developed. Bicarbonate concentration in the synthetic groundwater was similar to that found in UE-25 p#1. It was used primarily in experiments with uranium, neptunium, and plutonium because the solution and sorption behavior of these elements is sensitive to the bicarbonate and carbonate concentrations in solution. The derivation of sorption coefficient probability distributions for the elements of interest on the major rock types in Yucca Mountain involves both an evaluation of available experimental data and sorption modeling (BSC 2003e, Attachment I). Figure 4-1 shows an example of experimental and modeling results with significant variations in the sorption coefficients. The radionuclides of interest are divided into three groups of radionuclides. For the first group, including americium, neptunium, protactinium, plutonium, thorium, and uranium, experimental data are used to evaluate the impact of radionuclide concentrations, sorption kinetics, and variations in water chemistry on sorption coefficients. Surface complexation modeling is used to further evaluate the impact of variations in water chemistry and surface area on sorption coefficients. The surface complexation models used in this analysis are based on the code PHREEQC V2.3 (BSC 2001b). The impact of variations in pH on neptunium coefficients on devitrified tuffs is shown in Figure 4-1. The amount of scatter in the experimental data shows that the neptunium sorption coefficient depends very little on pH. PHREEQC surface complexation model curves reflect the equilibrium values of neptunium sorption coefficients on devitrified tuff (BSC 2003e, Attachment I). May 2004 4-2 No. 10: Unsaturated Zone Transport Revision 1 Source: BSC 2003e, Figure I-20. NOTE: PHREEQC model results for J-13 and synthetic UE-25 p#1 waters are also plotted. Figure 4-1. Neptunium Sorption Coefficient on Devitrified Tuff in J-13 and Synthetic UE-25 p#1 Waters versus Solution pH in Sorption (Forward) and Desorption (Backward) Experiments In the second group of elements, including cesium, radium, and strontium, the ranges of sorption coefficient values for the major rock types are derived directly from the available experimental data and the ranges for environmental variables expected in the transport system. Although it would be preferable to have a theoretical model to evaluate the impact of variations in water chemistry and rock chemistry on sorption coefficients for these radionuclides, sufficient data are not available to properly constrain such a model. For the third group, including carbon, iodine, and technetium, the sorption coefficient is set to zero in volcanic rocks and in alluvium. On the basis of the experimental and modeling results, the probability distribution function can be cumulative lognormal or truncated normal distributions. The distributions were constrained by pH, Eh, water chemistry, rock composition and surface area, and radionuclide concentration (see Section 5.3.2, Table 5-1, for Kds in the rocks of the unsaturated zone). In general, the approach used in derivation of the distributions tended to underestimate the range and median or expected value. This conservative approach was taken because of the potential scaling uncertainties in the application of these distributions to transport calculations at Yucca Mountain. May 2004 4.2 FIELD TESTS AND OBSERVATIONS Field tests were vital in the understanding of the processes and phenomena occurring during transport through the unsaturated zone. Even more importantly, they provided the basis for confidence building and validation of the numerical models used to predict the long-term fate and transport of radionuclides following potential releases at the repository. A summary of field experiments designed to investigate transport is presented in this section. These tests include the Busted Butte test series, the Alcove 8–Niche 3 test, and the Alcove 1 test. 4-3 No. 10: Unsaturated Zone Transport Revision 1 For the Alcove 1 test, numerical simulation and prediction were also presented here to provide implications of these field test results to model developments for the site-scale transport modeling. Transport tests that have been conducted up to now (and which are discussed here) involve length scales ranging from a few meters to a maximum of 30 m, and the time scale of these experiments does not exceed 2 years. There is a gap in the upper ranges of both the spatial and the temporal test scales, and a larger mountain-scale test or series of tests is practically infeasible. This gap impedes the application of the knowledge gained from the current set of field tests to the validation and support of the mountain-scale transport model. Nevertheless, the available test results are useful in establishing the basic conceptual framework of transport and for forming the basis for extension to large-scale transport. 4.2.1 Busted Butte Test Series The Busted Butte unsaturated zone transport test was a long-term experiment conducted in Busted Butte near Yucca Mountain (BSC 2003h, Section 6.13) to investigate flow and transport issues in the unsaturated zone process models for Yucca Mountain. The Busted Butte test facility is located in Area 25 of the Nevada Test Site, approximately 160 km northwest of Las Vegas, Nevada, and 8 km southeast of the repository area. The site was selected because of the presence of a readily accessible exposure of the Topopah Spring Tuff and the Calico Hills Formation, and the similarity of these units to those beneath the repository horizon (see Figure 4-2). The test facility consists of an underground excavation along a hydrogeologic contact between the TSw unit and the CHn unit. This contact is comprised of the nonwelded portion of the basal vitrophyre of the Topopah Spring Tuff. The study of the TSw–CHn interface is important because of the significant role that the vitric layers of the CHn unit play in radionuclide retardation (BSC 2003e, Section 6.10). The test proceeded in two phases, each differing in design, purpose, and experimental scales, among other factors. A detailed description of the tests can be found in Unsaturated Zone and Saturated Zone Transport Properties (BSC 2001c, Section 6.8; BSC 2003h, Section 6.13). The first phase, including test Phases 1A and 1B, was designed as a scoping study to assist in design and analysis of Phase 2 and as a short-term experiment aimed at providing initial transport data on a fracture near an interface. The second phase incorporated a larger region than Phase 1, with a broader, more complex scope for tracer injection, monitoring, and collection. 4.2.1.1 Phase 1A Test The Phase 1A Test was located in the CHn unit spanning both the geologic Calico Hills Formation (Tac) and the nonwelded subzone of the lowermost Topopah Spring Tuff (Tptpv1 in the lithostratigraphic units (BSC 2002); ch1v in the unsaturated zone hydrogeologic model layer (BSC 2003e). It involved the injection of nonreactive tracers into the Tptpv1 and the subsequent flow and transport in the Tptpv1 and Tac layers. It was a noninstrumented or blind test, consisting of four single-point injection boreholes. All Phase 1 boreholes were 2 m long and 10 cm in diameter. May 2004 4-4 No. 10: Unsaturated Zone Transport Source: BSC 2003h, Figure 6.13.1-1. NOTE: This schematic of the Busted Butte unsaturated zone transport test shows the relative locations of the different experiment phases and borehole locations. Figure not drawn to scale. ERT = electrical resistivity tomography. Figure 4-2. Busted Butte Unsaturated Zone Transport Test 4-5 No. 10: Unsaturated Zone Transport Revision 1 May 2004 Revision 1 Tracers were injected continuously for 286 days at rates of 1 mL/hr (boreholes 2 and 4) and 10 mL/hr (boreholes 1 and 3). The injected species included a mixture of conservative (nonsorbing) tracers (bromide, fluorescein, pyridone, and fluorinated benzoic acids), a sorbing tracer (lithium), and fluorescent polystyrene microspheres (i.e., an analog for a colloidal tracer). Of all the tracers injected during the field experiment, only the transport of the nonsorbing bromide and fluorescein tracers can be studied because of the poor quality and unreliability of the data from the other two tracers. A schematic of the borehole layout in the Phase 1A test is shown in Figure 4-3. More detailed NOTE: The figure is not drawn to scale. information is provided in Unsaturated Zone and Saturated Zone Transport Properties (BSC 2001c, Section 6.8). Only injection into borehole 3 (located in the Tptpv1 layer of the CHn unit) was considered. Borehole 3 is about 20 cm above the Tptpv1–Tac interface. The injection tests into boreholes 1 (located in the Tptpv1 layer, but further removed from the interface), 2, and 4 (located in the Tac layer of the CHn unit) were not analyzed because of the poor quality of the data. Source: BSC 2003h, Figure 6.13.1-3. May 2004 Figure 4-3. Phase 1A Borehole Locations The field test was completed through excavation by mineback and auger sampling. During mineback, as successive vertical slices were being removed, digital photographs under visible and ultraviolet light were taken to record the distribution of moisture and fluorescein. In addition, rock samples were collected by augering, and the exposed plane was surveyed. The auger samples were analyzed for tracer concentration. The fluorescein plume in the vicinity of borehole 3 at various locations along the y-axis (originating at the rock face and going into the rock) is shown in Figure 4-4. The plume cross sections show a relatively uniform distribution of fluorescein around the injection borehole, although some borehole shielding effects (tracer blocked or delayed from moving in the direction 4-6 No. 10: Unsaturated Zone Transport Revision 1 of the borehole) can be seen. At all of the mineback faces, the corresponding plume cross sections are more oval than round. Lithologic contacts clearly influence flow and tracer transport. Figure 4-4 shows the distinct geologic layering, denoted by the limited penetration and higher fluorescein concentrations in the less permeable lower layer. Source: BSC 2003h, Figure 6.13.2-1. 4.2.1.2 Phase 1B Test Phase 1B involved the injection of the tracers discussed in the Phase 1A test and collection of pore-water and tracer samples in the lower section of the Topopah Spring Tuff (Tptpv2). This test was designed to acquire data on fracture–matrix interactions in the TSw, providing some of the only such data. The results were used to calibrate fracture properties for the Phase 2 analysis. Figure 4-4. Fluorescein Plume at Each of Four Phase 1A Mineback Faces 4-7 May 2004 No. 10: Unsaturated Zone Transport Revision 1 In the Phase 1B field test, the tracers were injected into two boreholes (borehole 5 at a rate of 10 mL/hr, and borehole 7 at a rate of 1 mL/hr) in the lower portion of the Topopah Spring basal vitrophyre (Tptpv2 in lithostratigraphic units and tsw39 in the unsaturated zone layers of the hydrogeologic units), which is a relatively low-permeability fractured rock. Samples were obtained in collection boreholes 6 and 8 (Figure 4-5). Detailed information on the design, configuration, and dimensions of the Phase 1B test is provided in Unsaturated Zone and Saturated Zone Transport Properties (BSC 2001c, Section 6.8). The tracer solutions were injected at a depth of 1.30 m, measured from the rock face into the injection boreholes. Water samples from boreholes were collected and analyzed regularly during the injection period (Tseng and Bussod 2001). Source: BSC 2003h, Figure 6.13.1-4. NOTE: GPR = ground penetrating radar; ERT = electrical resistance tomography. Figure 4-5. Schematic of Phase 1B and Phase 2 Borehole Locations Breakthrough of all five solute tracers was observed in collection borehole 6, directly below the 10-mL/hr injection borehole 5. No tracer breakthrough was discernible in collection borehole 8 below the 1-mL/hr injection borehole 7. The breakthrough concentrations (the ratios of measured concentration, C to initial concentration, C0) of bromide and 2,6-difluorobenzoic acid (DFBA) in borehole 6 are shown in Figures 4-6 and 4-7, respectively. Maximum concentrations are invariably observed at a (horizontal) depth of approximately 1.3 cm into the rock face, which is directly underneath the injection port in borehole 5. However, the value of the maximum concentration and amount of mass recovery vary greatly. Bromide and 2,6-DFBA (both nonsorbing anionic tracers) exhibit reasonable breakthrough patterns and relatively similar maximum relative concentrations compared with other trackers. 4-8 May 2004 No. 10: Unsaturated Zone Transport Revision 1 Figure 4-6. Bromide Concentrations in Borehole 6 for Phase 1B Source: BSC 2003h, Figure 6.13.2-4a. Source: BSC 2003h, Figure 6.13.2-4b. Figure 4-7. 2,6-DFBA Concentrations in Borehole 6 for Phase 1B May 2004 4-9 No. 10: Unsaturated Zone Transport Revision 1 4.2.1.3 Phase 2 Tests Phase 2 tests were designed to incorporate large volumes of the rock. A detailed description of this test phase can be found in Unsaturated Zone and Saturated Zone Transport Properties (BSC 2001c, Section 6.8). Phase 2 testing involved a 7-m-high, 10-m-wide, and 10-m-deep block representing all of the rock units of Phase 1. The injection points for this phase were distributed in two horizontal, parallel planes arranged to test the properties of the lower Topopah Spring Tuff (Tptpv2) and the hydrologic Calico Hills (Tptpv1 and Tac). Six upper injection boreholes (boreholes 18 to 23) (of which only three were used) and four lower boreholes (boreholes 24 to 27) were drilled into the block, as shown in Figure 4-5. The upper injection plane consisted of 37 injection points distributed along the axes of the injection boreholes and was located in the fractured Topopah Spring Tuff (Tptpv2). As in Phase 1B, this unit represents the base of the TSw basal vitrophyre and is characterized by subvertical fractured surfaces that form columnar joints. The natural fracture pattern present in this unit serves as the conduit for tracer migration into the CHn. The lower horizontal injection plane was located in the Calico Hills Formation (Tac) and included 40 injection points distributed in the four horizontal and parallel injection boreholes. These boreholes were designed to incorporate the lower part of the block in the event that the top injection system did not involve the entire block during the testing program. Phase 2 also included 12 collection boreholes drilled in a direction perpendicular to that of the injection boreholes (see Figure 4-5, boreholes 9 to 17 and boreholes 46 to 48). These boreholes contained collection pads evenly distributed on inverted membranes to collect samples. Phase 2 included three subphases: 2A, 2B, and 2C. Phase 2A involved injection (at a rate of 1 mL/hr per injection point) into a single instrumented borehole (borehole 23) in the upper injection plane. This borehole is located entirely within the Tptpv2 unit, which consists of fractured, moderately welded tuffs from the basal vitrophyre. In Phase 2B, four instrumented injection boreholes (boreholes 24 to 27) in the lower injection plane were used, and the injection rate was much higher than in Phase 1A (10 mL/hr per injection point). In this case, the injection plane was restricted to the Calico Hills Formation (Tac). Thus, the Phase 2B test was designed to incorporate the lower section of the test block, while the upper section of the block was incorporated during the Phase 2A and 2C tests. Phase 2C involved injection into three upper boreholes (boreholes 18, 20, and 21) at much higher rates than in Phases 1A and 1B (50 mL/hr per injection point). As in Phase 1A, the injection system was located on a horizontal plane in the Tptpv2 unit. The tracers used in the Phase 2 tests included all the tracers used in Phase 1. Additionally, three fluorinated benzoic acids, a mixture of sorbing solute species (Ni2+, Co2+, Mn2+, Sm3+, Ce3+, and rhodamine WT), and a nonsorbing anionic tracer (I-) were used. Pad analyses confirm breakthrough of the nonsorbing tracers in 14 of the 15 collection boreholes. Of the sorbing tracers, breakthrough has been confirmed only in the case of lithium in 10 of the 15 collection boreholes. The spatial distributions of bromide and lithium in sampling borehole 16 at different times during Phase 2C tests are shown in Figures 4-8 and 4-9. The lower lithium concentrations are consistent with the difference in the sorption behavior of the two tracers. May 2004 4-10 No. 10: Unsaturated Zone Transport Source: DTN: LA0112WS831372.003. NOTE: BH = borehole. Figure 4-8. Spatial Distributions of Bromide in Sampling Borehole 16 at Different Times During Phase 2C Test Source: DTN: LA0201WS831372.007. NOTE: BH = borehole. Figure 4-9. Spatial Distributions of Lithium in Sampling Borehole 16 at Different Times During Phase 2C Test No. 10: Unsaturated Zone Transport Revision 1 May 2004 4-11 Revision 1 4.2.2 The Alcove 8–Niche 3 Test The drift-to-drift tests at Alcove 8–Niche 3 (in progress) provide information on seepage and transport over spatial scales in the range of 20 m. This is the intermediate scale for relating site-scale processes of seepage and percolation with drift-scale processes of diversion and seepage. Along the long flow paths, the corresponding advective transport is affected by fracture–matrix interaction, which is shown to be an important retardation mechanism that delays the movement of water and tracers through the unsaturated units. A detailed description of the Alcove 8–Niche 3 test can be found in In Situ Field Testing of Processes (BSC 2003h, Section 6.12). Figure 4-10 shows the location of the test site within the ESF main drift and the Enhanced Characterization of the Repository Block (ECRB) Cross-Drift, as well as a three-dimensional representation of the test area, including several slanted (near-vertical) boreholes. Alcove 8 is located within the upper lithophysal zone of the TSw (Tplpul). This unit has some lithophysal cavities that may intersect fractures. The distinctive feature of the test bed in Alcove 8 is a nearvertical fault that cuts across the floor. The fault is open on the ceiling of the alcove, and appears to be closed along the floor. In the test, water and tracers were released in trenches about 5 cm wide and about 5 cm deep that had been etched along this fault. Niche 3 (also referred to as Niche 3107) is approximately 4 m wide, extends to approximately 14 m from the centerline in the ESF main drift, and is about 20 m below the floor of the alcove. The niche is located within Tptpmn, about 20 m below the floor of the alcove; the interface between Tptpul and Tptpmn is about 15 m below the floor of Alcove 8 (BSC 2003h, Section 6.12). The fault in Alcove 8 is visible along the ceiling of Niche 3. Water was introduced along the fault (about 5 m long) (DTN: GS020508312242.001) under ponded conditions (with 2 cm of water head) until quasi-steady state seepage was observed in Niche 3. Then, a finite volume of water containing two tracers with different molecular diffusion coefficients (bromide and pentafluorobenzoic acid (PFBA)) was introduced into the fault. Once the tracer-laced water had been released into the fault, more tracer-free water was released. Both tracer-laced and tracer-free releases occurred under the same ponded condition. This release of tracer-free water continued until a few months after breakthrough of the two tracers (bromide and PFBA) was observed in the seepage collected in Niche 3. A water-pressure head of 2 cm was applied at a ponding plot along the fault in Alcove 8. The plot consisted of four trenches of different application rates as a result of heterogeneity along the fault. Figure 4-11 shows the total application rate as a function of time. The considerable temporal variability of the application rate observed during the test was attributed to heterogeneity in (1) the degree of infill and (2) in the properties of the infill materials within the fault just below the ponding plot (Figure 4-10) and to the unsaturated state of the majority of the fault and of the surrounding fractures away from the plot. After 209 days of water application, two nonsorbing tracers with different molecular diffusion coefficients (bromide and PFBA) and a sorbing tracer (lithium) were introduced into the water at the infiltration plot. Tracer concentrations were measured in three of the trays (at the niche) capturing seeping water from the fault. Figure 4-12 shows the evolution over time of bromide concentration and of the daily seepage rates for a 1.5-month period following arrival of the May 2004 4-12 No. 10: Unsaturated Zone Transport Revision 1 wetting front. The bromide concentration increases from a low initial level (about 3 ppm) to the release concentration of 30 ppm about 30 days after the onset of seepage. Given the nonsorbing behavior of bromide, the delayed bromide breakthrough is a good indicator of the importance of matrix diffusion, the absence of which would have resulted in much earlier observations of the 30 ppm concentration. Source: BSC 2003h, Figure 6.12.1-2. Figure 4-10. Test Bed for the Alcove 8–Niche 3 Tests May 2004 4-13 No. 10: Unsaturated Zone Transport Revision 1 Source: i BSC 2003e, F gure 7.4-1. Figure 4-11. Total Application Rate as a Function of Time Source: BSC 2003h, Figure 6.12.2-7. Figure 4-12. Concentration of Bromide Plotted against Seepage Rates Measured for 45 Days after First Observations of Drips in Tray 6 May 2004 4-14 No. 10: Unsaturated Zone Transport Revision 1 Figure 4-13 shows the tracer concentration in the seepage water at two sampling locations (tray 7 and tray 9+23, sampling trays in field tests, which were picked for discussion because they are the points where lower and upper seepage fluxes were observed). In tray 7, both bromide and PFBA were first detected three weeks after the initial release of the tracers into the fault. The concentration of both tracers gradually increased, though the rise in PFBA concentration clearly preceded that of bromide. This indicates a lower molecular diffusion coefficient for PFBA, and is consistent with the relative sizes of the two species. After peaking, the concentrations decreased and finally reached a relatively constant level. Similar patterns were observed in tray 9+23, but the magnitude and times at which peak concentrations occurred were different, suggesting the involvement of different fractures with different flow pathways, leading to different advective transport. This is further supported by the measured lithium concentrations, which are much higher in tray 9+23 (than that in tray 7) because the shorter pathway and faster flowing fractures intercepted in this case allow less time for matrix diffusion and retardation through sorption. Similar tracer behavior with different diffusion coefficients was observed by Reimus et al. (2003) in tracer experiments in fractured volcanic tuffs at the C-Wells site. This consistency suggests that similar transport processes (advection and matrix diffusion) are at work in the unsaturated and saturated zone barrier systems in the fractured tuffs of low matrix permeability. May 2004 4-15 No. 10: Unsaturated Zone Transport Source: BSC 2003h, Figure 6.12.2-8. NOTE: Cm is the measured concentration and Ca is the background concentration in the leachate. Figure 4-13. Concentration of Tracers Measured in Seepage in Niche 3 No. 10: Unsaturated Zone Transport Revision 1 May 2004 4-16 Revision 1 5. MODEL DEVELOPMENT This section provides a brief discussion of the sequence of events, processes, and phenomena involved in the transport of radionuclides through the unsaturated zone in Yucca Mountain following the release of radionuclides from the repository. This section introduces the corresponding conceptual and mathematical models, and describes the general simulation approach. Additionally, it describes model development and calibration using a series of field tests on progressively larger spatial and temporal scales. Model calibration demonstrates the ability of the conceptual and mathematical model to adequately capture the underlying transportrelated processes in the various tests, establishes confidence in the models, and forms the basis for model validation. 5.1 OBJECTIVES OF THE RADIONUCLIDE TRANSPORT STUDY The primary objectives of the radionuclide transport study are: 1. To integrate the available data for the development of a comprehensive model of radionuclide transport through the unsaturated zone of Yucca Mountain for a range of current and future climate conditions 2. To identify the controlling transport processes and phenomena, to determine their relative importance, and to evaluate the effectiveness of matrix diffusion and sorption as retardation processes 3. To identify the geologic features that are important to radionuclide transport 4. To estimate the migration of important radionuclide solutes and their decay products from the repository toward the water table 5. To evaluate the effects of various climatic conditions on radionuclide transport 6. To estimate the migration of radioactive colloids from the repository toward the water table, and to determine the sensitivity of colloid transport to the kinetic coefficients of colloid filtration 7. To evaluate through sensitivity analyses the effect of uncertainty in important parameters on the predictions of transport. These objectives are met through the activities discussed in Sections 5 and 6. 5.2 INTEGRATION OF INFORMATION IN THE DEVELOPMENT OF TRANSPORT MODELS The development of conceptual models of radionuclide transport through the unsaturated zone involves the synthesis of information from various sources. The corresponding mathematical models are based on inputs stemming from these sources, including studies on hydrology (as affected by the climate), geology and stratigraphy, and radionuclide transport properties and May 2004 5-1 No. 10: Unsaturated Zone Transport Revision 1 parameters. This section discusses the interaction of these components and conditions with transport processes. The processes that control the radionuclide migration through the unsaturated zone are: (1) advection, which is the transport of radionuclides by flowing water; (2) dispersion and diffusion, which lead to a wider distribution of radionuclides over larger portions of the unsaturated zone domain; and (3) sorption of solutes or filtration of colloids, which leads to attachment onto solid grains and removes radionuclides from flowing water. The latter process enhances both diffusion and retardation. All are separately discussed in Section 3. To understand the integrated transport phenomena in the unsaturated zone system, it is imperative to identify processes, properties, conditions, and their interactions. Conditions that are essential for unsaturated zone transport are: 1. Precipitation, infiltration, and deep percolation–Precipitation, infiltration and deep percolation, as these are affected by the climatic conditions, describe a critical source term of transport (i.e., the amount of water that is expected to enter the top boundary of the unsaturated zone). Spatial and temporal variations in the infiltration and percolation rates have a direct impact on water flow and saturation patterns in the unsaturated zone. Under the conditions of fracture-dominated flow, prevalent in the unsaturated zone, the effect on radionuclide advection through the unsaturated zone is immediate, greatly influencing the radionuclide residence time in the unsaturated zone. Additionally, the variations in the saturation patterns in the fractures and the matrix affect diffusive transport (specifically matrix diffusion) and dispersion. Predicting percolation fluxes also requires an understanding of climatic conditions through studies of historical climate and precipitation patterns and their projection into the future. Data from these analyses determine the range of expected precipitation rates during the compliance period. These precipitation rates are then used as the basis for estimating net infiltration, which is the key factor affecting percolation rates and flow patterns and, thus, transport behavior. 2. Stratigraphy and hydrogeologic properties–The unsaturated zone system geology, stratigraphy, and hydrogeologic properties define flow pattern of the percolating water in the unsaturated zone. Hydraulic conductivity, wettability, porosity, and fracture characteristics and properties of the various hydrogeologic units, in addition to their fracture characteristics and properties, determine the ability of the unsaturated zone to conduct and store water and are responsible for the existence of fast flow pathways (usually associated with conductive faults and interconnected vertical fractures), perched water bodies (when more permeable units overlie less permeable ones), as well as flow focusing and lateral diversion. Information on the system geology and hydraulic properties is obtained by integrating data from: (1) visual observations from outcrops and the ESF; (2) geophysical logging; (3) core data analysis from boreholes at select locations; (4) test data obtained in the course of pneumatic and/or water injection field tests; and (5) inversion of the field data. May 2004 5-2 No. 10: Unsaturated Zone Transport Revision 1 Correlating information of the system geology and stratigraphy (including identification of the main hydrogeologic units, their areal extent and thickness, and the presence of faults) with the corresponding rock hydraulic properties and integration into a comprehensive flow model of the unsaturated zone allows the determination of flow fields for the infiltration scenarios discussed in Item 1. For the reasons discussed in Item 1, the flow field and the corresponding water saturation distribution have a direct and significant impact on radionuclide transport. 3. Source terms–Radionuclide transport is directly affected by the source term, which describes the combined releases of radionuclides and of water (i.e., the transport agent). Water contacting the waste packages provides the main transport mechanism for radionuclide releases from the repository. Seepage through the ceiling of the waste emplacement drift determines the amount of water that can potentially contact waste. If no seepage occurs, the mobilization and release of radionuclides will be impeded. By accounting for the range of possible climatic scenarios and through the determination of seepage-relevant parameters, the range of seepage rates (and the corresponding effect on transport) can be estimated. Seepage is affected by small-scale processes and occurs at a scale smaller than that of mountain-scale precipitation and the corresponding flow fields. Consequently, seepage estimates were obtained from experimental and modeling work performed on the drift scale (information on the subject can be found in Technical Basis Document No. 3: Water Seeping into Drifts). In general, water is diverted away from the waste by vaporization (during the thermal period), diversion around the drift due to the capillary barrier effect, the drip shield, and the waste package. These protective measures and mechanisms would all have to fail at the same location along the waste emplacement drift for water to contact the waste. Upon release from the package, radionuclide-contaminated water enters the invert at the base of the drift. The extent of radionuclide sorption by the invert material affects release rates to the unsaturated zone. The effect of the invert (which consists of crushed tuff) on sorbing radionuclide species is to reduce the radionuclide source term for unsaturated zone transport. Chemical reactions leading to precipitating species or colloid flocculation can also reduce the radionuclide source term. Information on radionuclide source reduction due to species immobilization is provided by the study of near-field geochemistry. If the released radionuclides do not sorb, or sorb and escape the invert in colloidal form, the water carries them to the large pore space between the crushed tuff blocks of the invert. 4. Solutes versus colloids–The form of the released radioactive species (dissolved or as colloids) affects the release mechanism and release rate. If the water contacts the contents of the waste package, then radionuclides are released at the base of the drift. The speciation of the mobilized radionuclides determines the magnitude of matrix diffusion (a function of the molecule/colloid size and electrical properties), sorption May 2004 5-3 No. 10: Unsaturated Zone Transport Revision 1 onto (and filtration by) the unsaturated zone rocks, and pore-size exclusion (for colloids only). Information on the nature of the released radionuclides and their concentrations in the water (the species-related component of the source term) is provided by knowledge of the radioactive substances stored in the waste packages and analysis of phase and speciation changes occurring as a result of chemical reactions and radioactive decay. Additionally, these studies provide information on the expected release concentrations, completing the definition of the radionuclide source term. 5. Shadow zone–Because of capillary pressure barrier effects, a shadow zone with reduced saturation and reduced flux is likely to be created underneath the drift. The presence of the shadow zone can lead to significant retardation of radioactive solute and colloid transport (BSC 2003m) by reducing the amount of fracture flow and significantly lowering water saturations relative to the surrounding rock. Because of the low advective flux in the shadow zone, diffusive flux is the only mechanism capable of transporting radionuclides from the invert to the shadow zone and past the boundaries of the shadow zone. The shadow zone has a greater retardation effect on colloid transport (compared to solute transport) because of the much lower diffusion coefficient of colloids. 6. Transport through the unsaturated zone below the repository–Once released from the invert, radionuclides are carried by water from the invert’s large pores toward the deeper, undisturbed unsaturated zone (fracture advective transport), toward which a small amount of radionuclides is also migrating very slowly through the invert matrix (diffusion and matrix advective transport). The transport of these radionuclides to the water table through the unsaturated zone is significantly affected by the flow field below the repository (through its direct impact on advection), as determined by the interaction of climatic, geologic, source, and near-field conditions discussed in Items 1 through 6. Transport to the water table is also determined by matrix diffusion and the sorption (for solutes) or filtration (for colloids) behavior of the radioactive species. Matrix diffusion depends on the diffusion coefficients of the various species. This information is available from literature sources or is obtained from laboratory experiments (see Section 4.1). Additional factors affecting matrix diffusion are the fracture frequency and characteristics (obtained from the field and laboratory studies discussed in Item 2 and inversion exercises with the measurements) and the distributions of tortuosity, porosity, and water saturation in the various rocks of the unsaturated zone profile (also available from the studies in Item 2). Information on the sorption behavior is provided from field and laboratory studies, as discussed in Section 4. Conventionally, an effective linear equilibrium sorption model is assumed with a sorption Kd of the species in question onto the solid phase, including the various unsaturated zone rocks, and natural pseudocolloids, such as clays (in the case of colloid-assisted radionuclide transport). Such field or laboratory studies have been limited to sorption onto the rock matrix. No information is available for the May 2004 5-4 No. 10: Unsaturated Zone Transport Revision 1 estimation of sorption onto fracture walls. In the absence of such data, a conservative approach is followed in the models, in which no sorption or filtration is considered in fractures. Field and laboratory tests are also necessary for the determination of colloid filtration in the unsaturated zone rocks. While limited studies are available, applicability to the unsaturated zone is not well established because the studies were performed under fully water-saturated conditions. Colloid pore-size exclusion (straining, a potentially significant mechanism for larger colloids) can be estimated from pore-size distributions, which can be estimated from the capillary pressure curves of the unsaturated zone rocks under consideration based on field data from Item 2. In light of this uncertainty, the approach followed in the models involves analysis of transport using a very wide range of the corresponding parameters, thus bracketing the possible solutions. 7. Fracture versus matrix flow and fracture–matrix interaction–Because of the fracture-dominated flow patterns indicated by the studies in UZ Flow Models and Submodels (BSC (2003b, Section 6.6), radionuclides that enter the undisturbed unsaturated zone underneath the repository tend to move rapidly toward the water table through a continuous fracture network. Advection through the matrix would occur and lead to further radionuclide arrivals at the water table but at a far slower rate. The main mechanisms of retardation are diffusion into the matrix through the fracture walls and through the matrix, and sorption (for solutes) or attachment-filtration (for colloids) onto the matrix, which further enhances retardation by increasing diffusive fracture–matrix fluxes. Radioactive decay is a mechanism to reduce the mass of radionuclides arriving at the water table. Sorption onto natural pseudocolloids, such as clays, can be important in colloid-assisted solute transport, if the natural colloids exist in sufficiently large concentrations. Information on the subject is provided by the study of colloids at the site (BSC 2003d). Decay species can have significantly different diffusion and sorption behavior. For example, the 235U daughter of 239Pu has a far lower affinity for sorption, leading to faster transport of daughter radionuclides to the water table than that predicted by accounting solely for the parent. Thus, the contribution of daughters significantly impacts the transport of total radionuclides to the water table. 8. Transport in the saturated zone–After traveling through the unsaturated zone, radionuclides arrive at the saturated zone. The study of flow and transport through the unsaturated zone provides the water and radionuclide source terms, respectively, for the saturated zone studies. These are also based on data obtained from laboratory and field studies on the saturated zone hydrology and transport behavior, which is documented in Technical Basis Document No. 11: Saturated Zone Flow and Transport. May 2004 5-5 No. 10: Unsaturated Zone Transport Revision 1 5.3 CONCEPTUAL AND NUMERICAL MODEL The development of the conceptual and numerical model constitutes an integral component of the study to meet the objectives presented in Section 5.1. The model of radionuclide transport considers a large three-dimensional mountain-scale domain. The fractured rock is conceptualized as a heterogeneous dual-permeability system, in which the distinct hydraulic and transport behavior of fractures and matrix is described by using separate properties and corresponding parameters. This conceptualization allows the description of the complex unsaturated zone flow field (the most important aspect of the transport study), in which fracture flow plays a dominant role. The grid, conditions, and calibrated hydraulic parameters used in the transport simulations are identical to those used for the analysis of flow in UZ Flow Models and Submodels (BSC 2003b, Section 6). These parameters correspond to the permeability barrier model (called the conceptual model of perched water), which uses the calibrated perched-water parameters for fractures and matrix in the northern part of the model domain, and modified property layers (including tsw38, tsw39, ch1z, and ch2z model layers), where the lower basal vitrophyre of the TSw overlies the zeolites of the CHn. A detailed discussion of this perched-water model can be found in UZ Flow Models and Submodels (BSC 2003b, Section 6.2). A two-dimensional plan view of the grid at the repository level is shown in Figure 5-1. All three-dimensional transport simulations are based on the steady-state flow fields from UZ Flow Models and Submodels (BSC 2003b, Section 6). The numerical process model used for simulating transport (BSC 2003e, Sections 4 and 6.2) accounts for all major known transport processes. Those processes are: (1) advection; (2) molecular diffusion; (3) hydrodynamic dispersion; (4) kinetic or equilibrium physical and chemical sorption (linear, Langmuir, Freundlich, or combined model); (5) radioactive decay and tracking of resultant daughter radionuclides; (6) colloid filtration (equilibrium, kinetic, or combined); and (7) colloid-assisted solute transport. Data for the description of the sorption behavior of the investigated radionuclides are taken from the analysis of laboratory experiments (see Section 4.1), while a wide range of possible filtration parameters is used to compensate the limited availability of experimental data. Experimentally determined data (either directly available in the literature or scaled to account for relative size and behavior) are used for the diffusion coefficient, D0, of solutes (Lide 1993, pp. 5-111 to 5-112), while the Stokes-Einstein equation (Bird et al. 1960, p. 514) is used to estimate the colloidal diffusion coefficients. 5-6 May 2004 No. 10: Unsaturated Zone Transport Revision 1 Figure 5-1. Plan View of the Unsaturated Zone Model Grid at the Repository Level Source: BSC 2003e, Figure 6.7-1. May 2004 5-7 No. 10: Unsaturated Zone Transport Revision 1 Radionuclides Considered.The following radioactive solutes were considered in the unsaturated zone transport model: . 99Tc (a nonsorbing species) . 237Np, 235U, and 233U (moderately sorbing species) . 241Am, 239Pu, 231Pa, 229Th, 226Ra, 90Sr, and 135Cs (strongly sorbing species). Their properties are listed in Tables 5-1 and 5-2. The radionuclides are selected to represent the groups that have different transport magnitudes of sorption, molecular diffusion, and radioactive decay. Due to the lack of such information for fission products of those radionuclides, the fission products are excluded from the transport study. The radionuclides are released at the grid blocks corresponding to the location of the repository (blue dots indicate the location of repository grids in Figure 5-1). Additionally, for the three-dimensional simulations of continuous release, all the important members in the decay chains of 241Am and 239Pu are considered, according to the decay equations (Pigford et al. 1980): 241 (Eq. 5-1) Am ¡æ 237Np ¡æ 233U ¡æ 229Th 239 (Eq. 5-2) Pu ¡æ 235U ¡æ 231Pa Only the most important members of the radioactive chain are included in these decay equations, which omit decay products with short half-lives because they have a minor effect on the relative abundance of the daughters. The transport of radioactive colloids was also simulated. Spherical PuO2 colloids with diameters of 6, 100, 200, and 450 nm were used in this transport model. May 2004 5-8 No. 10: Unsaturated Zone Transport Table 5-1. Kd in the Rocks of the Unsaturated Zone Species Uranium Neptunium Plutonium Americium Protactinium Cesium Strontium Radium Thorium Unit Zeolitic Devitrified Vitric Zeolitic Devitrified Vitric Zeolitic Devitrified Vitric Zeolitic Devitrified Vitric Zeolitic Devitrified Vitric Zeolitic Devitrified Vitric Zeolitic Devitrified Vitric Zeolitic Devitrified Vitric Zeolitic Devitrified Vitric Source: BSC 2003e, Table 6.5-1. No. 10: Unsaturated Zone Transport Kd value (mL/g) 0.5 0.2 0.2 0.5 0.5 1 100 70 100 500 1,000 400 10,000 10,000 10,000 5,000 7.5 2 1,000 40 25 2,500 500 300 15,000 5,000 5,000 Revision 1 Ranges of Kd value (mL/g) (Kd value, probability) (0, 0) (0.5, 0.5) (30, 1.0) (Kd value, probability) (0, 0) (0.2, 0.5) (4, 1.0) (Kd value, probability) (0, 0) (0.2, 0.5) (3, 1.0) (Kd value, probability) (0, 0) (0.5, 0.5) (6, 1.0) (Kd value, probability) (0, 0) (0.5, 0.5) (6, 1.0) (Kd value, probability) (0, 0) (1.0, 0.5) (3, 1.0) (Kd value, probability) (10, 0) (100, 0.5) (200, 1.0) (Kd value, probability) (10, 0) (70, 0.5) (200, 1.0) (Kd value, probability) (10, 0) (100, 0.5) (200, 1.0) Range = 100 to 1000 (500) Range = 100 to 2000 (1,000) (Kd value, probability) (100, 0) (400, 0.5) (1,000, 1.0) Range = 1000 to 20,000 (10,000) Range = 1000 to 20,000 (10,000) Range = 1000 to 20,000 (10,000) (Kd value, probability) (425, 0) (5,000, 0.5) (20,000, 1.0) Range = 1 to 15 (7.5) (Kd value, probability) (0, 0) (2, 0.5) (100, 1.0) Range = 50 to 2000 (1000) Range = 10 to 70 (40) Range = 0 to 50 (25) Range = 1000 to 5,000 (2,500) Range = 100 to 1,000 (500) Range = 50 to 600 (300) Range = 1,000 to 30,000 (15,000) Range = 1,000 to 10,000 (5,000) Range = 1,000 to 10,000 (5,000) May 2004 5-9 Table 5-2. Radionuclide Properties in the Transport Simulations T1/2 Radionuclide 99Tc 237Np 239Pu 241Am 233U 235U 231Pa 229Th D0 (m2/s) 4.55 × 10-10 1.65 × 10-10 4.81 × 10-10 3.69 × 10-10 4.94 × 10-10 4.89 × 10-10 4.98 × 10-10 5.02 × 10-10 8.89 × 10-10 2.06 × 10-9 (years) 2.13 × 105 2.14 × 106 2.41 × 104 4.322 × 102 1.59 × 105 7.08 × 108 3.25 × 104 7.90 × 103 1.599 × 103 2.30 × 106 2.90 × 101 226Ra 135Cs 90Sr 7.91 × 10-10 Source: BSC 2003e, Table 6.5-2. 5.4.1 The Busted Butte Phase 1A Test The Busted Butte test was used to predict tracer concentration and water saturation following the injection of nonreactive tracers into the Ttptv1 and the subsequent flow and transport in the Tptpv1 and Tac layers. The Busted Butte test series is described in Section 4.2.1, and the Phase 1A test is described in Section 4.2.1.2. The transport of two (nonsorbing tracers Br- and fluorescein) is analyzed in the transport model report (BSC 2003e). In this analysis, the fluorescein data were used for calibration while the Br- data were used to increase confidence in the model. For this three-dimensional numerical study, the underlying geologic model considered a homogeneous and anisotropic unfractured rock matrix with the properties of the Tptpv1 and Tac layers. Because of the injection configuration (as described in Section 4.2.1), only half the domain (i.e., the portion of the domain to the right of the injection point; see Figure 5-2) was simulated, using a grid consisting of 22,463 elements (BSC 2003e). Although the configuration of injection did not lead to a symmetrical system, the effect of such asymmetry was expected to be limited and confined to the earlier part of the field test. TOUGH2 V1.1 MEOS9nT V1.0 (LBNL 1999) was used for the simulations. 5.4 MODEL DEVELOPMENT AND CALIBRATION The process of calibration determines the main hydraulic parameters of the medium under consideration. The calibrated hydraulic parameters are then used to increase confidence in the transport model using transport parameters available from other independent sources. Corroboration is based on comparison with experimental data for the Busted Butte tracer transport (Sections 5.4.1 to 5.4.3) or other field tests (Alcove 8–Niche 3 test, Section 5.4.4). Calibration is conducted during model development to enhance confidence in the radionuclide transport model. 5-10 No. 10: Unsaturated Zone Transport Revision 1 May 2004 Revision 1 Figures 5-2 and 5-3 show the fluorescein distribution (as recorded digitally in the field using an ultraviolet light) and the corresponding numerically predicted distribution (i.e., calibrated) at a distance of y = 0.90 m mineback face. The determination of the hydraulic parameters is a significant component of the calibration effort. A comparison of the two figures indicates good agreement between predictions and observations. The absence of stringers and other features indicative of fracture presence in the field data, coupled with the ellipsoidal size of the plume, supports the assumption of an unfractured (porous) medium. Visual evidence of local heterogeneity in Figure 5-2 notwithstanding, the plume is quite uniform at the scale of observation. Although the simulated horizontal dimension of the plume is somewhat larger than the observed one, the prediction accurately captures several important features of the observed plume: its ellipsoidal shape, the compression of the lower half of the plume that results from its proximity to the less permeable Tac, and the vertical dimension of the fluorescein plume. Due to data deficiencies, calibration cannot be further refined. NOTE: The correct scale is the faint feature (20 cm) in the middle of the photograph below the plume. Source: BSC 2003h, Figure 6.13.2-2. Figure 5-2. Fluorescein Plume at the Mineback Face (y = 90 cm) at Borehole 3 5-11 May 2004 No. 10: Unsaturated Zone Transport Source: BSC 2003e, Figure 7.3-2. Figure 5-3. Numerical Prediction (Relative Concentration, C/C0) of the Fluorescein Plume Using Calibrated Parameters (Busted Butte Test Phase 1A) No. 10: Unsaturated Zone Transport Revision 1 May 2004 5-12 Model confidence is increased by using the calibrated properties obtained in the fluorescein simulation (Table 5-3) and the known diffusion coefficient (D0) of Br. to predict the transport of bromide. Figure 5-4 shows the numerically predicted Br. distribution (relative concentration, C/C0) at the y = 0.90 m mineback face and the field measurements of the Br. relative concentration, which were obtained from samples taken during the mineback. Parametera ƒÓ kx (m2) ky (m2) kz (m2) ƒÑ ƒ¿ (1/m) c n c Ch2v 0.360 8.20 ~ 10-13 2.82 ~ 10-13 3.64 ~ 10-14 0.12 0.741 1.200 0.07 Sr Ch1v 0.320 2.14 ~ 10-13 4.14 ~ 10-13 6.28 ~ 10-14 0.22 0.471 1.332 0.07 4 ~ 10-10 D0 of Fluorescein (m2/s) Table 5-3. Calibrated Parameters of Flow and Transport from the Analysis of the Busted Butte Phase 1A Field Test May 2004 Source: BSC 2003e, Table 7.7-3. The overall predicted and measured Br. distributions agree. The model reasonably reproduces the rather uniform concentrations along the x-axis that crosses the injection point, as well as the magnitude of the observed concentrations. Some discrepancies between observations and predictions appear toward the outer extent of the horizontal axis of the plume, where concentration gradients are steep, and, therefore, differences may be magnified. However, the agreement between observations and numerical predictions is sufficiently good to enhance the confidence of the radionuclide transport model. Uncertainty in the measurements may also affect the match between prediction and observations. The measured relative concentration of 2.77 (with respect to the injection concentration) in the center of the plume may be real (if, for example, the sample was inadvertently allowed to dry), but it may also be indicative of some error in the measurements. 5-13 No. 10: Unsaturated Zone Transport Revision 1 Revision 1 Source: BSC 2003e, Figure 7.3-3. NOTE: The solid circles indicate the location of measurements (DTN: LA9910WS831372.008), which appear in the corresponding boxes. Figure 5-4. Field Measurements and Numerical Prediction (Relative Concentration, C/C0) of the Bromide Distribution in Busted Butte Test Phase 1A May 2004 5-14 No. 10: Unsaturated Zone Transport Revision 1 5.4.2 Busted Butte Phase 1B Test The Busted Butte Phase 1B test is described in Section 4.2.1.3. Of the five tracers injected during Phase 1B, measurement data for 2,6-DFBA (assumed to be nonsorbing but demonstrating behavior akin to sorption after long times) and Br- (nonsorbing) were selected from the first 100 days of the Phase 1B test. The 2,6-DFBA data were used for calibration, and the Br- data set was used to increase model confidence. The initial (uncalibrated) input parameters of flow and transport for the simulation of the Phase 1B test were also provided by the same sources as in the Phase 1A test. The same model assumptions on model design and the grid used in the simulation of Phase 1A were used here. A comparison of the measured and numerically predicted breakthrough curves of 2,6-DFBA (based on the calibrated parameters, which include a small Kd value to describe the apparent sorption behavior at long times) in Figure 5-5 shows very good agreement. The shape and breakthrough time of the curves are consistent with the visual observation that the system did not exhibit fracture flow behavior during the Phase 1B test (despite the known presence of a fracture intersecting the two boreholes) and support the validity of the unfractured medium approach used in the simulations. Source: BSC 2003e, Figure 7.3-4. Figure 5-5. Observed and Numerically Predicted (Calibrated) Breakthrough Curves of 2,6-DFBA for the Busted Butte Phase 1B Test May 2004 5-15 No. 10: Unsaturated Zone Transport The nonzero (although small) distribution coefficient (Kd = 1.47 ~ 10.5 m3/kg) for 2,6-DFBA appears to contradict the conventional assumption that 2,6-DFBA is nonsorbing but is actually supported by the long-term sorption behavior of the tracer. In this case, some sorption is plausible because of the nature of 2,6-DFBA (an organic acid) and because of the very low permeability of the rock, which leads to long residence times of the solute. Organic substances are generally considered to be slow-sorbing (Cameron and Klute 1977). Additionally, the long residence time in the rock may allow a slow reaction of 2,6-DFBA with minerals in the rock, a process that would result in a response consistent with an apparent sorption. The calibrated parameters (based on the 2,6-DFBA transport analysis) were used for the prediction of Br. transport in Phase 1B of the Busted Butte test (Table 5-4). The known D0 of Br. and its Kd value of zero were used to increase confidence in the simulation. A comparison of the measured and numerically predicted breakthrough curves of Br. in Figure 5-6 shows very good agreement. The level of agreement between observations and numerical predictions enhance the confidence of the radionuclide transport model. Table 5-4 Parametera ƒÓ kx (m2) ky (m2) kz (m2) ƒÑ tsw39 0.270 3.06 ~ 10-17 4.06 ~ 10-17 1.53 ~ 10-17 0.07 1.47 ~ 10-5 Kd of 2,6-DFBA (m3/kg) Calibrated Flow and Transport Parameters from the Analysis of the Busted Butte Phase 1B Field Test May 2004 Source: BSC 2003e, Table 7.3-5. 5-16 No. 10: Unsaturated Zone Transport Revision 1 Revision 1 Source: BSC 2003e, Figure 7.3-5. 5.4.3 Busted Butte Phase 2C Test The Busted Butte Phase 2 tests are described in Section 4.2.1.4. The analysis for this model development and calibration activity is limited to Phase 2C, which appears to have yielded the best quality data. Because of the relatively short spacing between injection boreholes and the long injection period (695 days), the injection can be assumed to be uniform along the three boreholes (see Figure 4-5, boreholes 18, 20, and 21). The implication of this uniformity is that it is possible to model the system using a two-dimensional grid, thus allowing higher spatial resolution. After reviewing the tracer concentrations recorded at the collection boreholes, the analysis concentrated on data from borehole 16 because it is the closest to the horizontal plane of the three injection wells and, thus, registered the strongest signals (in terms of tracer concentrations). Borehole 16 extends along a horizontal plane (roughly coinciding with the Tptpv2–Tptpv1 interface) about 0.6 m below the plane of the injection boreholes and perpendicular to their main axis. Based on a review of concentration data for all of the tracers injected, two tracers, Br- and Li+, were selected for calibration and model confidence building. The use of Br- and Li+ offered two advantages: they registered relatively strong and almost always consistent signals in borehole 16, and they were injected as a solution of LiBr, with equal numbers of Li+ and Br- moles in the Figure 5-6. Observed and Numerically Predicted Breakthrough Curves of Bromide in the Busted Butte Phase 1B Test May 2004 5-17 No. 10: Unsaturated Zone Transport Revision 1 system, which provided an additional mass balance constraint. In this analysis, the Li+ data were used for calibration and the Br. data for model confidence. Preliminary simulations indicated that matrix diffusion and Li+ sorption would be needed to describe the joint transport of Br. and Li+. This required that the model domain be a fracture. matrix system, as opposed to a porous medium (without fractures) or one that considered transport only in fractures. A dual-permeability model was employed with a fine-resolution two-dimensional grid (.x = .y = 0.05 m). The initial flow field was obtained by setting an infiltration rate corresponding to 5 mm/yr at the top of the domain and running the flow simulation to steady state. This had to be done at every phase of the model development and calibration process because the changes in the hydraulic parameters during the process affected the flow field. For the Li+ calibration process, only the data for times of 337 and 440 days were used. Because of the sorbing behavior of Li+, the data for times of fewer than 337 days were marked by very low concentrations, significant variability, and corresponding uncertainty. Concentration data for times of greater than 440 days showed rather inconsistent behavior attributed to issues discussed in Section 7.3 of Drift-Scale Radionuclide Transport (BSC 2003i). Br. for 125 and 183 days were used to increase confidence during model development. The data for times less than 125 days were marked by very low concentrations and significant variability. Concentration data for times greater than 183 days indicated a rather uniform distribution along the borehole (as the concentration peak had already been reached). Figure 5-7 shows the measured and the numerically predicted Li+ distributions (based on the calibrated parameters shown in Table 5-5) along the collection borehole. The general features of the predicted Li+ distribution agree with the measured distribution. The locations of the peaks, the peak concentration values (except for the left-most peak), and the general concentration distributions are matched. Figure 5-8 shows the measured and numerically predicted Br. distribution along the collection borehole. This distribution was obtained by using the calibrated properties from the analysis of the Li+ data, the known D0 and zero sorption for Br. . The level of agreement between observations and numerical predictions enhances the confidence of the radionuclide transport model. The distribution of Br. indicates a less diffusive behavior relative to Li+, because Br. is nonsorbing. The presence and location of concentration peaks (both observed and predicted) are consistent between the two tracers. May 2004 5-18 No. 10: Unsaturated Zone Transport Source: BSC 2003e, Figure 7.3-6. Figure 5-7. Observed and Numerically Predicted (Calibrated) Breakthrough Curves of Li+ in the Busted Butte Phase 2C Test Table 5-5. Calibrated Flow and Transport Parameters from the Analysis of the Busted Butte Phase 2C Field Test Parametera ƒÓ kx (m2) kz (m2) 10-13 8 ~ 10-14 7.1 ~ 10-13 F-TpTpv1 1 3 ~ 10-13 2 - M-TpTpv1 0.354 1.3 ~ 10-13 0.9 5.5 ~ 10-4 - ƒÑ Kd of Li (m3/kg) Kd of Li (m)b 2.5 ~ 10-6 Source: BSC 2003e, Table 7.3-8. 5-19 No. 10: Unsaturated Zone Transport Revision 1 M-TpTpv2 0.060 1.2 ~ 10-13 8 ~ 10-14 0.654 9.3 ~ 10-4 - F-TpTpv2 1 1.96 ~ 10-12 2 - 4.3 ~ 10-6 May 2004 Revision 1 Source: BSC 2003e, Figure 7.3-7. Phase 2C Test Inspection of the simulated Br- distribution reveals that the numerical solution accurately predicts the location and magnitude of the concentration peaks but exhibits narrower peaks and indicates a deeper trough between the second and third peaks compared to the measurements. More specifically, the measured concentration of Br- is more uniform along the collection borehole axis than what the numerical simulation predicts, indicating a system that is more diffusive than the advective one described by the simulation. The discrepancy between the prediction and measured data is attributed to the effects of the collection pad on transport through the host rock. The properties of the collection pad, characterized by high permeability, porosity, irreducible water saturation, and capillary pressure, differ significantly from those of the host rock. Tracer-carrying fracture flow first reaching the collection pad arrives at distinct points (the points where the fractures intercept the borehole) and is quickly redistributed on the initially dry pad. Therefore, the tracer in the pad is more uniform than the one in the overlying rock and the samples indicate a more diffusive system. Under these conditions, it is expected that the most accurate data will be at early times and at locations roughly un