A/ K/\ x

/>

THREE-DIMENSIONAL MODELING OF UNSATURATEDFLOW IN THE VICINITY OF EXPLORATORY SHAfl ,FACILITIES AT YUCCA MOUNTAIN'/ NEVADA' /

M. L. RockholdB. SagarM. P. Connelly(a)

December 1989

Paci fic Northwest-'tboratoryRichland4 Washington 99352

(a) 'Westinghouse Hanford CompanyRichland, Washington

"I/

':10rn0

0

-3Mz

zC-31WM

GO

N

EXECUTIVE SUMMARY

This report was prepared by Pacific Northwest Laboratory (PNL),ffor the

Office of Civilian Radioactive Waste Management (OCRWM) of the Department of"'

Energy (DOE). The study was undertaken to estimate the moisture distribution

of unsaturated fractured media in the vicinity of an exploratory/,-shaft'facility

at the proposed high-level nuclear waste repository site neatucca"MQuntihC

Nevada.

BACKGROUND AND OBJECTIVE /

The exploratory shaft facility (ESF) forms an importarttpart of the

subsurface-based site characterization activities that are outlined in the

Yucca Mountain Site Characterization Plan (SCPYK{DOE 1988). During the

construction and operation of the ESF, methods would-be selected to minimize

rock disturbances. However, some alterations in the properties of rocks close

to the excavations are to be expected. The objective of this study is to

investigate the influence of these alte ed-ock zones on moisture distribution

in the vicinity of the ESF.

SCOPE

This study is limited to investigation of either extremely long-term

(steady-state) or short-term (on the order of 20 yrs) impacts of the ESF on

the vadose zone moist*e distribution. Any \l'ong-term effects are important

in the estimation of waste migration and fate. Any short-term effects may be

important in the/planning and interpretation of tests performed in the ESF.

METHOD

The PORFLO-3 computer code (Runchal and Sagar 1989; Sagar and Runchal

1989) is used for unsaturated flow simulations. Rather than represent fractures

as disc ete ments, an 'equivalent continuum' is stipulated in which the

fraciured-units are assigned equivalent or composite hydrologic properties.

Explicit treatment of fractures is not feasible because of the extremely large

number of fractures contained in the site-scale problem and the difficulties

in characterizing and modeling the fracture geometries. Three-dimensional

geometry is. used for the simulations. The exploratory shafts and drifts and

the altered property rock zones (called the Modified Permeability Zones (MPZs]

in the context of hydrology) associated with them are represented as

iii

one-dimensional line elements that are embedded in the general three-dimensional

elements. The Ghost Dance Fault, one of the prominent structural and

hydrological features is also included in the calculation domain. This fault--

is represented as a two-dimensional planar element.

CONCLUSIONS N

The results of these simulations suggest that the modi yper;eAjlit

zones (MPZs) around the exploratory shafts may act as conduits for preferential

flow of water in the welded tuff units when locally ,yaturated conditions and

fracture flow occur. However, these simulations also suggest that fracture

flow, as represented in this model with equivalent c6ntiniaum Approximations,

is limited to the Tiva Canyon welded tuff unit.. The Topopah Springs weldedtuff unit is sufficiently thick so that most episodic pulses of water from

natural precipitation should not reach the repository through fractures due

to capillary imbibition into the rock matrix. These conclusions, however,

are specific to the initial and boundar conditions, and material properties

used for these simulations. A potential for sigificant lateral flow above the

interface between the Paintbrush nonweldediuvit and thk Topopah Springs welded

unit is also indicated. Where laterally flowing water intersects preferential

flow paths or structural features such as fault zones, locally saturated or

perched water table conditions could develop which could lead to fracture

flow in the Topopah Springs welded unit (ie.,'the repository horizon). The

Ghost Dance Fault was embedded within the model domain for these simulations,

but the properties used to describe'it are estimates due to a lack of data.

The simulations also suggest.that-there will be no appreciable change in the

moisture-content distribution below the Tiva Canyon welded unit within 20

years after the construction of. the ESF. Therefore, even though excavations

associatecwith the ESF will~alter the fracture densities and permeabilities

of the tuff units. the ambient moisture contents and distributions in the

vicinity of the shafts should not change significantly within a reasonable

time frame for subsurface characterization work.

RECOMMENDATIONS

Post-closure thermal effects and vapor-phase transport of moisture caused

by heat generated from the repository were not considered in this study.

However, these factors may have a significant affect on the performance of

iv

the repository. Therefore, future post-closure modeling studies should consider

non-isothermal flow and vapor-phase transport of moisture. The MPZs and shaft

backfill material may become more important for future smaller-scale simulations

when coupled fluid flow, heat transfer, and mass transport are considered.

Future simulations with intermediate and larger scale multi.dimensipnal

models of the proposed repository at Yucca Mountain should cosiider thetopography of the site and the locations of surface expressions of major\'- '

fractures and faults to determine potential areas forpreferential local /recharge. Then, spatially non-uniform, time-varying recfiarge fluxes should

be applied as surface boundary conditions to obtal'n the most-accurate

representations of flow fields in the natural systenfi-atdepth.

v

CONTENTS

EXECUTIVE SUMMARY . . . . . . . . . . . .

1.0 INTRODUCTION . . . . . . . . . . . .

2.0 MODEL DESCRIPTION . . . . . . . . . .

2.1 GEOMETRY OF CALCULATION DOMAIN

2.2 HYDROLOGIC PROPERTIES

2.3 INITIAL CONDITIONS . . . . . . .

2.4 BOUNDARY CONDITIONS . . . . . .

3.0 COMPUTER CODE DESCRIPTION . . . . . .

4.0 RESULTS AND DISCUSSION . . . . . . .

4.1 STEADY-STATE SIMULATIONS . . . .

4.2 TRANSIENT SIMULATIONS ... .

5.0 CONCLUSIONS AND RECOMMENDATIONS . . .

6.0 REFERENCES . . . . . . . . . . . . .

21.1

2.1

. .

. . ..t .. .. . 2.15

t .. .. . . . 2.17

. . . .

3.1

4.1

4.1

4.13

5.1

6.1

vi

FIGURES

1.1 Reference Location of the Perimeter Drift of the ProposedRepository at the Yucca Mountain Site . . . . . . . . . . . . . 1.2

1.2 Conceptual Illustration of the Exploratory Shaft Facility'. . . 1.3

2.1 Vertical Dimensions used for PORFLO-3 Simulations . . L./. * . *

2.2 Yucca Mountain PORFLO-3 Computational Grid . . . . . . . . . . . Z. 3

2.3 Relative Location of the Model Domain . . . . . . . . . . . . 2.4

2.4 (a) Matrix Hydraulic Conductivity Curve for Unit PTn . . . . . 2.8

(b) Relative Hydraulic Conductivity and 'aturation Curves forUnit PTn . . . . . . . . . . . . . . . . . . . . . . . . . . 2.8

2.5 (a) Matrix Hydraulic Conductivity Curve for Unit Chnv . . . . . 2.9

(b) Relative Hydraulic Conductivity and Saturation Curves forUnit Chnv ................ . 2.9

2.6 (a) Composite Hydraulic Conductivity-Curve for-Unit TCw . . . . 2.10

(b) Relative Hydraulic Conductivity and Saturation Curves forUnit TCw . . . . . . . . . . . . . . . . . . . . . . . . . . 2.10

2.7 (a) Composite Hydraulic Conductivity Curve for Unit TSwl . . .. 2.11

(b) Relative Hydraulic Conductivity and Saturation Curves forUnit TSw1 . . .. 2.11

2.8 (a) Composite Hydraulic Conductivity Curve for Unit TSw2-3 . . . 2.12

(b) Relativeflydraulic Conductivity and Saturation Curves forUnit TSw2-3 . . . . . . . . . . . . . . . . . . . . . . . . 2.12

2.9 (a) Hydraulic Conductivity Curve Representing Shafts and Drifts-i-a-Unit TSw2-3.. . 2.15

(b),-Relative\-Hydraulic Conductivity and Saturation CurvesRepresenting Shafts and Drifts in Unit TSw2-3 . . . . . . . 2.15

4.1 (a) Steady-State Relative Saturation Profile from the Center ofthe Model Domain for the Case 1 Simulation (without Shaftsand MPZs) . . . . . . . . . . . . . . . . . . . . . . . . . 4.3

4.1 (b) Steady-State Pressure Distribution from the Center of theModel Domain for the Case 1 Simulation (Initial ConditionsSpecified for Case 2 simulation) . . . . . . . . . . . . . . 4.3

vii

4.2 (a) Steady-State Relative Saturation Profile from the Center ofthe Model Domain for the Case 2 Simulation (with Shafts andMPZs) .. -. 4.5

4.2 (b) Steady-State Pressure Distribution from the Center of theModel Domain for the Case 2 Simulation (with Shafts andDrifts) . . . . . . . . . . . . . . . . . . . . . . . 4.5

4.3 (a) Steady-State Relative Saturation Profile from theCenter of-,,the Model Domain for the Case 2 Simulation . . . . . . . . . 4.6

4.3 (b) Steady-State Pressure Distribution through Line ElementsRepresenting ES-1 for the Case 2 Simulation .. . . . . . . 4.6

4.4 Vertical Cross-Section of Darcy Velocity Vectors/through-Plane Intersecting ES-1 for the Case 2 Simulation .4.8

4.5 Horizontal Cross-Section of Darcy Velocity Vectors through theLower Part of Unit TCw for the Case 1 Simulation . . . . . . . . 4.9

4.6 Horizontal Cross-Section of Darcy Velocity Vectors through theLower Part of Unit TCw for the Case 2 Simulation . . . . . . . . 4.9

4.7 Steady-State Pressure Distribution ini'Horizontal Cross-Sectionthrough Lower Part of Unit TCw for-the Case 2 Simulation(corresponding to Figure 4.6) .4.10

4.8 Steady-State Relative Saturations in Horizontal Cross-Sectionthrough Lower Part of Unit TCw for the Case 2 Simulation(corresponding to Figure 4.7). . . . . . . . . . . . . . . 4.11

4.9 Horizontal Cross-Section of Darcy Velocity Vectors throughthe Upper Part of Unit TSx2-3 for the Case 1 Simulation . . . . . 4.12

4.10 Horizontal Cross-Section of Darcy Velocity Vectors throughthe Upper Part of Unit TSx2-3 for-the Case 2 Simulation .4.12

4.11 Steady-State Pressure Oistribution in Horizontal Cross-Sectionthrough Upper Part of Unit TSw2-3 for the Case 2 Simulation . . . 4.14

4.12 Steady!-State Relative Saturation in Horizontal Cross-Sectionthrough Upper Part of Unit TSw2-3 for the Case 2 Simulation'(correspondihtg to Figure 4.11) .... . . . . . . . . . . . . . 4.15

4A13 Horizontal Cross-Section of Darcy Velocity Vectors throughtheProposed Repository Horizon in Unit TSw2-3 for the Case 1Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.16

4.14 Horizontal Cross-Section of Darcy Velocity Vectors throughthe Proposed Repository Horizon in Unit TSw2-3 for the Case 2Simulation ..... . . . . . . . . . . . . . . . . . . . . . . 4.16

viii

4.15 Steady-State Pressure Distribution in Horizontal Cross-Sectionthrough the Proposed Repository Horizon in Unit TSw2-3 for theCase 2 Simulation (corresponding to Figure 4.14) . . . . . . . . 4.17

4.16 Steady-State Relative Saturations in Horizonal Cross-Sectionthrough the Proposed Repository Horizon in Unit TSw2-3 for,theCase 2 Simulation (corresponding to Figure 4.15) . . . . 4.18

4.17 Relative Saturation Profiles for Unit TCw from the Cente'rKofthe Model Domain for the Transient Simulations . . . . . . . . 4. 4.20

4.18 Relative Saturation Profiles for Unit TCw along thi PlaneRepresenting the Ghost Dance Fault for the Transient

- Simulations . . . . . . . . . . . . . . . . .... . . . . . . . . 4.21

4.19 Relative Saturation Profiles through Line Elements RepresentingES-1 in Unit TCw for the Transient Simulations . . . . . . . . . 4.22

TABLES

2.1 Fracture Characteristics of Hydrologic Units . . . . . . . . . . 2.5

2.2 Parameters for Matrix Hydrologic Properties . . . . . . . . . . . 2.6

2.3 Saturated Hydraulic Conductivity Multyplying Factors for theModified Permeability Zone . . . . . . . . . . . . . . . . . . . 2.14

2.4 Initial Conditions for-the Case 1 Steady-State Simulation . . . . 2.17

ix

1.0 INTRODUCTION

The geologic formations in the unsaturated zone underlying Yucca Mountain-,

Nevada, are being evaluated by the U.S. Department of Energy as/thepotential

location for a high-level radioactive waste repository. Investigptions are

currently underway to evaluate the hydrologic conditions, prR ses A ndproperties of the unsaturated zone at this site.

Yucca Mountain consists of a series of North-trending fault-block ridges

composed of volcanic ash-flow and ash-fall tuffs that generally have a regional

dip of 5 to 7 degrees to the East (Scott and Bonk 1984). Some of the welded

tuff units are highly fractured, which requires the&hydrologic properties of

these units to be evaluated before estimating the-rate at'which radionuclides

could migrate to the accessible environment. Determining the significance of

fracture flow at Yucca Mountain is important because the estimated composite

saturated hydraulic conductivities of the welded tuff units are up to 3 orders

of magnitude greater than the saturated hydraulic conductivities of the matrix

in the welded tuffs. Therefore, if fracture flow is significant, it could

greatly reduce the travel time of water percolating-downward through the

repository horizon. The proposed repository area, depicted in Figure 1.1, is

also bounded by steeply dipping faults or by fault zones, and is transected

by a few normal faults,-Therefore, the effects of fault zones on unsaturated

water flow and contaminant transport should also be investigated.

The exploratory shaft facility '(gSF) forms an important part of the

subsurface-based site characterizatijon activities that are outlined in the

Yucca Mountain Site Characterization Plan (SCP) (U.S. Department of Energy

1988). The ESF will consist of surface facilities and underground excavations

as depicted-in the conceptuaV picture shown in Figure 1.2. The underground

excavations include, two 4.4 m diameter shafts (called ES-1 and ES-2) connected

by drifts at the proposed repository horizon. Descriptions of 34 test

activities to be performed in the ESF are provided in the SCP. While the

construction and operation methods will be selected to minimize rock

disturbances, some alterations in properties of rocks close to the excavations

are to be expected. Therefore, it is important to determine the effects of

1.1

1%*IiII

SOUTHERN NEVADA

'I 1

NI '-

le NAFR 1.L

'- L -..1

tV-/

I..Iz

. 0

LAS VEGAS

N765000-

N76000

N755000

a

kI

X.40 0 40 80 Km i

25 0 25 50 miles

i a- REFERENCE LOCATION OFUNDERGROUND REPOSITORY

. _ ..T.. ... .... .F ....

I. .. , . ... ... - I-.--.iCON~TOUR INTERVAL t00 FEET

YU CCAE 560000 E 565000

MOUNTAIN SITE

FIGURE 1.1. Reference Location of the Perimeter Drift of the ProposedRepository at the Yucca Mountain Site (PBQ and D, 1985,GE/CALMA, 1986, Product No. 0119)

1.2

EXP VES STORAGEAREA

ES-2

ES-1 .

f TTO ORI .-HOLE WASH

UPPER"DEMONSTRATION

BREAKOUT ROOM- LEVEL|

TO IMBRICATEFAULT

DANCE FAULT

FIGURE 1.2.. Conceptual Illustration of the Exploratory Shaft Facility

1.3

these altered-rock zones, also referred to as modified permeability zones(MPZs), on the moisture distribution in the vicinity of the ESF.

This study is limited to investigation of either extremely long-term

(steady-state) or short-term (on the order of 20 years) impacts/-of the ESF onthe vadose zone moisture distribution. Any long-term effects/are-important

for estimating the migration and fate of wastes. Any short4Aym effec~ts m

be important in the planning and interpretation of the tests performed inmthe

ESF. The effects of the ESF on the long-term (on the Order of 10,000 yrs) >-

post-closure performance of the repository are also important but do not forma part of this study.

In order to study the spatial effects of the ESF, three-dimensional

geometry is used in the simulations. The exploratory shafts and drifts and

MPZs associated with them are represented as one-dimensional line elements

that are embedded in the general three-dimensional elements. The Ghost Dance

Fault, one of the prominent structural-and hydrological features is also

included in the calculation domain. This fault-is represented as a two-

dimensional planar element. Results of both steady-state and transientsimulations are presented.

1.4

2.0 MODEL DESCRIPTION

2.1 GEOMETRY OF CALCULATION DOMAIN

For the purpose of these simulations, the lithologic units at Yucca

Mountain have been grouped into 5 hydrologic units which are,'deplcted'in

Figure 2.1. From top, these units are: the Tiva Canyon welde/unit"RCw) >the

Paintbrush Tuff non-welded unit (PTn), the Topopah Springs welded unit 1"'(TSw1j,

the Topopah Springs welded units 2 and 3, which are treated as a single

composite unit (TSw2-3), and the Calico Hills non-welded unit (CHnv). The

respective thicknesses of these units are 25, 40, 130, 205, apd 130 meters.

The relative elevations of the various units are shown in Figure 2.1 (after

Peters et al. 1986). These relative elevations are based on the stratigraphy

of drill hole USW-G4, which is located near the eastern edge of the prospective

repository boundary. The decision to treat TSw2 and TSw3 as a single composite

unit (TSw2-3) is based on the similarity of the hydrologic properties of the

two units and the relatively small thicknes~s-of TSw3 (15 m) relative to that

of TSw2 (190 m). The proposed repository will be located within unit TSw2

at a depth of approximately 310 m below the ground surface (from drill hole

USW-G4).

Computations are performed in a three-dimensional Cartesian grid system.

The axes of this system are aligned with East-West (X-axis), North-South (Y-

axis) and vertical (Z-axis). The dimensions of the calculation domain are

615 m, 300 m, and 530 m in'the X-, Y-4 and Z-directions, respectively. A total

of 25,056 nodes (29 in X-, 16 in Y-< and 54 in Z-direction) are used to

discretize this domain for numerical calculations. This grid is shown in

Figure 2.2. The Ghost Dance Fault, which is located near the West boundary,

is represented with a vertical planar element (two-dimensional) extending

from ,the~ground surface to the water table. The exploratory shafts (ES-1 &

ES-2) are represented as vertical line elements (one-dimensional) extending

from the surface to a depth of approximately 310-m into the repository horizon.

The shaft-elements are connected by offset horizontal line elements,

representing connecting drifts, at the 310-m depth in unit TSw2-3. The

exploratory shafts and drifts have a diameter of 4.4 m and are assumed to

have zones of modified permeability around them. For the simulations reported

2.1

530 m (Ground Surface)TCw 5

- 505 m/

PTn - 465 m "

TSw1 i

- 335m

TSw2-3

130-m

CHnv

- 0.0 m (Water Table)

FIGURE 2.1. Vertical Dimensions used for PORFLO-3 Simulations.TCw - Tiva Canyon welded unit; PTn - Paintbrush Tuffnon-welded unit; TSwl,2,3 - Topopah Springs weldedunits; CHnv - Calico Hills non-welded, vitric unit

here, the MPZs are assumed to extend 1 m past the edges of the shafts and

drifts. Case and KelsalI (1987) report the value of 1 m as an upper bound

estimate of-the MPZ dimension.

The shafts and fault are positioned within the model domain with reference

to borehole USW-G4 (Scott and Bonk 1984; Fernandez et al. 1988). The relative

location of the model domain is shown in Figure 2.3. The geometry of the

computational grid was designed in consideration of the spatial position of

the exploratory shafts relative to the Ghost Dance Fault. The line of symmetry

dividing the North and South halves of the model domain roughly corresponds

2.2

ES-2

Ghost Dance Fault

TiCw

PTn

CHn v

tab I*

FIGURE 2.2. Yucca Mountain PORFLO-3 Computational Grid. TCw - TivaCanyon welded unit; PTn - Paintbrush Tuff non-welded unit;TSw1,2,3 - TopopahNSprings welded units; CHnv - CalicoHills non-welded, vitric unit

with the position of Coyote Wash which is an East-West trending waterthed

that drains off of Yucca Ridge, located to the West. The intersection of

Coyote Wash-with the Ghost Dance Fault is a potential location for preferential

local recharge.

2.2 HYDROLOGIC PROPERTIES

The welded tuff units (TCw, TSw1, and TSw2-3) are distinguished hydro-

logically from the nonwelded units (PTn and CHn) by their observed fracture

density. Table 2.1 lists the fracture characteristics of the five tuff units.

2.3

Scale 1:1 2,000Contour Interval 20 feet

o 500 / \.|We t e ris

0 1000 Fe e

FIGURE 2.3. Relative Location of Model Domain. (Base Map fromScott and Bonk 1984)

2.4

TABLE 2.1. Fracture Characteristics of Hydrologic Units(after Peters et al. 1986)

Fracture Fracture FractureSample Aperture Density(b) Fracture Conductivity

Unit Code(a) (microns) (no./m 3) Porosity(c) (m/s)

TCw G4-2F 6.74 20 1.4E-4 3. 3.8E-05PTn G4-3F 27.00 1 2.7E-5 6.1E-M04TSw1 G4-2F 5.13 8 4.1E-5 2.2E-05TSw2-3 G4-2F 4.55 40 1.8E-4 1.7E-05CHnv G4-4F 15.50 3 4.6E-5 2.OE-04

(a) Fracture properties were measured on a limited number of samples usingconfined fracture-testing methods described by Peters et al. (1984).The differences in the fracture properties of the 3 welded units (TCw,TSW1, TSW2-3) reflect different confining pressures representing overburdenweight evaluated at the average unit depth in USW-G4.

(b) Based on a report by Scott et al. (1983).(c) Fractures are assumed to be planar, therefore, fracture porosity is

calculated as fracture volume (aperture times 1 square meter) times numberof fractures per cubic meter.,

The hydrologic properties of the rock matrix-and the fractures may differ

considerably. The parameters of the water retention and hydraulic conductivity

functions (van Genuchten 1978; Mualem 1976) used to represent the rock matrix

of the five hydrologic units are displayed in Table 2.2 (after Peters et al.

1986). The corresponding-parameters for the fractures are estimated by Wang

and Narasimhan (1986) to be: Sr = 0.0395, a = 1.2851/m, n = 4.23. These

fracture parameters were used in equivalent continuum approximations of

hydrologic properties to represent the fractures in all 3 welded units.

The van Genuchten (1978) water retention curve is defined by:

Se = Sr + (Ss - Sr)[1 + (a#)n]-m; m = (1 - 1/n) (2.1)

where Se = effective or relative saturation

Sr = residual saturation

Ss = total saturation (F 1)

# = capillary pressure head

a, n, m = are empirical parameters.

2.5

TABLE 2.2. Parameters for Matrix Hydrologic Properties(after Peters et al. 1986)

Grain HydraulicSample Density Conductivity

Unit Code (g/cm3) Porosity (m/s) Sr ((1/rn) n

TCw G4-1 2.49 0.08 9.7E-12 0.002 8.21E-3 1.S58PTn GU3-7 2.35 0.40 3.9E-07 0.I00" 4.50E-2KX.872TSw1 G4-6 2.58 0.11 1.9E-11 0.080 5.67E-3 1.798TSw2-3 G4-6 2.58 0.11 1.9E-11 0.080 5.67E-3 1.798 ,'CHnv GU3-14 2.37 0.46 2.7E-07 0.041 1.60E-0 3.872

The effective porosity, ne, of each unit is calculated as

ne = nb(1 - Sr), (2.2)

where nb is the bulk porosity listed in Table 2.2.

In this analysis, rather than represent-the fractures as discrete

elements, an 'equivalent continuum' is stipulated in-which the fractured units

are assigned equivalent or composite properties. Explicit treatment of the

fractures is not feasible because of the extremely large number of fractures

contained in the site-scale problem and the difficulties in characterizing

and modeling the fracture geometries.

In the equivalent continuum approximation, the fracture porosity, Of, can

be defined as the ratio of the fracture volume to the total bulk rock volume.

The matrix porosity, Om, can be defined as the ratio of the volume of the

voids in the matrix (excluding fracture void space) to the volume of the matrix.

A bulk, equivalent porosity, Ob. can then be defined after Nitao (1988) as,

-b = Of + (1 - Of)#m (2.3)

Given the relative saturations of the fractures, Sf, and matrix, Sm. the

equivalent bulk saturation, Sb, can be defined as

Sb = Sfif + SM(l - Wom (2.4)

2.6

The weighting procedure of Klavetter and Peters (1986) is used to obtain the

equivalent bulk hydraulic conductivity,

Kc = Km(l - #f) + Kfff, (2.5)

where Kf, Km, and Kc are the fracture, matrix, and composite conductivities'

respectively.

Because the two nonwelded units have relatively few fractures, their

composite properties are approximately equivalent to those of the rock matrix.

Thus, for the Paintbrush Tuff (PTn) and Calico Hills (CHnv) nonwelded units,

the properties listed in Table 2.2 are used directly'. Graphical representations

of the properties of these two units are shown in figures 2.4 and 2.5. In

Figures 2.4a and 2.5a, hydraulic conductivity is plotted as a function of

pressure head. In Figures 2.4b and 2.5b, relative hydraulic conductivity and

pressure head are shown as functions of saturation.

Figure 2.6a shows the composite hydraulic conductivity curve for the

Tiva Canyon welded (TCw) unit. A large jump in the value of the composite

hydraulic conductivity at a certain value of pressure head (a -1 m in Figure

2.6a) is a distinguishing characteristic of the composite curve. In physical

terms, it indicates that at a certain tension (or equivalently at a certain

value of moisture content), flow switches from the matrix-dominated to fracture-

dominated flow and vice versa. This abrupt change in the hydraulic conductivity

introduces extreme nonlinearity in the problem which can lead to numerical

instabilities. Figure 2.6b shows the composite relative conductivity and

pressure head as functions of saturation for unit TCw. Figures 2.7 and 2.8

show the composite hydrologic properties of the TSwl and TSw2-3 units of the

Topopah Sp-rings unit. In the numerical model, to accurately represent the

abrupt changes in-conductivity caused by fracture flow, the hydrologic

properties of units TCw, TSwl, and TSw2-3 are introduced in a tabular form

(rather than the analytic van Genuchten relation).

Calculations based on these composite properties assume that there are

enough fractures of varied orientation to ensure isotropic behavior. These

calculations also assume each unit to be homogeneous except where hydrologic

properties have been modified for the shafts, drifts, MPZs, and the Ghost Dance

2.7

I

I E-~5

IE-6

I E-7

- w -

C

(I-,

E 1E-8

*5 1E-9Iz

:3-a 1E-10

0N)

;O U IE-11

a-" IE-12

IE-13

.. ... 1. .. . .... ... ...1 .. ....1 . .1 ..j

1E6

1 E5

1 E4

E 1000

o 100

I

L. 10

1E-1I- E 2

1

1E-1

1E-2

1 E-3

I1E- 4

1E-5 ;uCD

1E-6 0

1E-7 CD

C-)IE-8 a

1E-9 'C)

IE-1 `

1E-12

1E-13

IE--14

1E-15

1 E-1 6

IE-14

1E-15 1E-3

1E-31E-21E-1 1 10 100 1000 1E4

Pressure Head (-m)

FIGURE 2.4a. Matrix Hydraulic Conductivity Curvefor Unit PTn

0.0 0.2 0.4 0.6 0.8 1.0

Saturation

FIGURE 2.4b. Relative Hydraulic Conductivity andSaturation Curves for unit PTn

1E-5

1E-6

1 E-7

(n

E 1E-8

-4-j

*5; 1E-9

-a 1E-10C:

0to l E-11

L-1E-12

tr1E-13

1E-1 4

I-I

. ... .... . . - i , 1 .. ... J . . . -

1E6

1E5

1E4

-a0

L-

0~

1000

100

10

1

1

1E-1

1E-2

1E-3

lE-4

1E-5 ;Uo

1E-6 0

1 E-7 CD

lE-8 0

1E-9 L0

1E-10 c

1E-11

1E-12

1E-13

1E-14

lE-15

1E-16

1E-1

1 E-2 f.1E-3 I * I - I

0.0 0.2 0.4 0.6

Saturation

. .. I

1E-3 1E-

..... ..... ..... .... . - ...... ....... _ ...... ....... .. . , .......

-2 lE-1 1 10 100 1000 1E4

Pressure Head (-m)

0.8 1.0

FIGURE 2.5a. Matrix Hydraulic Conductivity Curvefor Unit Chnv

FIGURE 2.5b. Relative Hydraulic Conductivity andSaturation Curves"for nit Chnvi

''I

1E-5

1E-6

1 E-7

E 1E-8

1 1E-9 -

- lE-lO0

1 jlE-1 1 .

,I lE-12.

1E-13

1E-15 .

1E-31E-21E-1 1 10 100 1000 1E4

Pressure Head (-m)

FIGURE 2.6a. Composite Hydraulic ConductivityCurve for Unit TCw

1E6

1 E5

1E4

1000

E 1000I

a. 100

I--

0 100

Q)

UF

0~ L- 1:

1E-1

1 E-2

1E-3

1

1E-1

1E-2

I E-3

IE-4

IE-5 ;UCD

1E-6 a

1E-7 CD

1 E-8 0

1E-9 C-

1E-12

1E-13

1E-14

1E-15

0.0 0.2 0.4 0.6 0.8 1.0

Saturation

FIGURE 2.6b. Relative Hydraulic Conductivity andSaturation Curves for nit TCw

/

lE-5 1E6 1

1E-6

1 E-7

1E-8

'5 1E-81-

4-

C-2t5 l E-lO

I-&.

1E-1

-o E-10

(j1E-151

i

I

r

.. ...... . . .

lE5

1E4

E 1000I

a 100

ZIL 1 0

U,

G)EL

1 E- 1

1 E-2

I E-3

1E-1

1E-2

1E-3

1E-4

1E-5 ;UD

1E-6 0

1E-7 CD

1E-8 0

1E-9 c0

1E-10 c

1E-11 I

1E-12

1E-13

1E-14

1E-15

1 E-1 6

1E-31E-21E-1 1 10 100 1000 1E4 0.0 0.2 0.4 0.6 0.8 1.0

Pressure Head (-m) Saturation

FIGURE 2.7a. Composite Hydraulic ConductivityCurve for Unit TSwl

FIGURE 2.7b. Relative Itydraulic Conouctivity andSaturation Curves for Unit TSwl!

'. , . j

1E-5

1E-6

1E-7

C')

E 1E-8

*5 1E-9

U

-o lE-10c0

o lE-11

-a

-n lE-12

IlE-1 3

1E6

1E5

lE4

E 1000

_0o 100

Q)

L 10

V)

0~

lE-1

lE-2

lE-1

1E-2

1E-3

1E-4

P3

1E-5 a(D

1E-6 0

1E-7 CD

CT,1E-8 0

1E-9 c0

lE-10 c

IF-1Il '<

1E-12

IE-14 r..

1E-3 1E-2 1E-1

1E-13

1E-1 4

IE-15

1E-1 61E-3

1 10 100 1000 1E4 0.0 0.2 0.4 0.6 0.8 1.0

Pressure Head (-im) Saturation

FIGURE 2.8a. Composite Hydraulic ConductivityCurve for Unit TSw2-3

FIGURE 2.8b. Relative Hydraulic Conductivity andSaturation Curves for'Unit TSw2-3

Fault. Isothermal conditions are also assumed, although the PORFLO-3 code will

solve non-isothermal flow and transport problems as well.

The air-entry pressure heads (1/a in the van Genuchten model) for the MPZs

associated with the shafts and drifts in each unit are assumed to be 0.5 times

as great as the air-entry heads of the rock matrix in each unit. This'-,

modification is somewhat arbitrary since no data are availab1e/but reflects>

the assumption that the excavation process will change the effective pore-7,

size distribution in the MPZs such that the MPZs will effectively behave as-,-

coarser porous media than the undisturbed rock matrix. The saturated hydraulic

conductivities of the MPZs are assumed to be 40 and 80 times"'greater than

those of the undisturbed matrix in units TSw1 and TSw2-3, respectively. These

multiplying factors are based on the upper-bound estimates reported by Case

and Kelsall (1987) as shown in Table 2.3. The saturated hydraulic conductivity

of the MPZs in unit TCw was also assumed to be 80 times greater than the matrix

hydraulic conductivity of this unit. This assumption was based on the

similarity of fracture densities and porosities between unit TCw and unit

TSw2-3 as shown in Table 2.1. For the nonwelded units (PTn and CHnv), the

saturated hydraulic conductivities of the MPZs were-assumed to be 20 times

higher than the hydraulic conductivities of the matrix of each unit. The

TABLE 2.3. Saturated Hydraulic Conductivity Multiplying Factors for theModified Permeability Zone(a) (after Case and Kelsall 1987)

Depth(m) Stress Redistribution_ Expected(b) Upper Bound(c)Without Blast Damaqe Case Case

Elastic Elastop lastic

100 15 20 20 40310 -- 15 40 20 80

(a) Equivalent permeability is averaged over an annulus 1 radius wide aroundthe 4.4 m (14.5 ft.) diameter exploratory shaft.

(b) This is based upon an elastic analysis with expected strength, insitustress, sensitivity of permeability to stress, and a 0.5 m wide blastdamage zone.

(c) This is based upon an elastoplastic analysis with lower bound strength,upper bound stress, greatest sensitivity of permeability to stress, anda 1.0 m wide blast damage zone.

2.13

non-welded units have a much higher porosity and lower fracture density than

the welded units. Therefore, it was assumed that permeability modifications

resulting from shaft construction in these units would be less extensive than

for the welded units. All other properties of the MPZs in each unit are assumed

to be equal to the properties of the undisturbed rock matrix in-the units.

The hydraulic properties of the shafts and drifts are riepresented withl

parameters describing the properties of a coarse material such as gravel,,

The van Genuchten (1978) model retention curve parameters used to represents-,

the shafts and drifts are: Sr = 0.0, a = 1.0 and n = 6.2. Figure 2.9 shows

these curves graphically. These water retention characteristics create a

capillary barrier which effectively restricts flow into the shafts and drifts

until the surrounding MPZs are close to saturation. Each section of the shafts

and drifts is assigned a saturated hydraulic conductivity value 1000 times

greater than the saturated hydraulic conductivity of the MPZ adjacent to it.

Thus, the rate at which water flows through the shaft and drift elements is

governed by the hydraulic conductivity of--the adjacent MPZs.

No data regarding the hydrologic properties of the rubblized shear zone

around the Ghost Dance Fault are presently available. Therefore, the fault

MPZ was arbitrarily modeled as a 0.5-m wide zone with air-entry pressure heads

0.5 times as great as the air-entry pressure head of the rock matrix in each

unit and saturated hydraulic conductivities 40 times greater than the

undisturbed matrix saturated hydraulic conductivities of each unit.

2.3 INITIAL CONDITIONS

Two steady-state simulations were performed. In the first, or Case 1,

the relative saturation in each unit was assumed to be uniform. Saturation

percentages determined from core samples from various units at Yucca Mountain

have been eported by Montazer and Wilson (1984). From these reported average

saturation-values, capillary pressure heads were calculated for input as initial

conditions to Case 1 simulations, using the matrix water retention curve para-

meters listed in Table 2.2. These average saturation values and initial

pressure heads for the various hydrologic units are listed in Table 2.4. The

shafts, drifts, and MPZ subdomain elements were excluded from the Case 1

simulation. The Case 1 simulation was performed to establish the initial

conditions for Case 2.

2.14

fn

E

I--,

0

C-

U

*0

1E-5

1E-6 _

1E-7

1E-8

1E-9

lE-10

1E-1 1

1E-12

IE- 13

1E-1 4

1E-15 #' aglm11E-3 1E-2 1E-1

1E6

1E5

1E4

E 1000

o 100

L 10Q)(a

Ca

L. 1 0D

1E-n

1E-2

1E-1

1E-2

1 E-3

1E-4

1E-5 .{U

1E-6 5-

1E-7 CD

1E-8 o

1E-9 CC

1E-10 O

1E-1i1 '

1E-12

1E-13

1E-14

1E-15

NJ

1E-3 1E-16

1 10 100 1000 1E4 0.0 0.2 0.4 0.6 0 .a 1.0

Pressure Head (-m) Saturation

FIGURE 2.9a. Hydraulic Conductivity Curve Repre-senting Shafts and Drifts in UnitTSw2-3

FIGURE 2.9b. Relative Hydraulic ConduCtivity andSaturation Curves Representing Shaftsand Drifts in Un it W2-3

K~

TABLE 2.4. Initial Conditions for the Case 1 Steady-State Simulation(saturation values from Montazer and Wi1son 1984)

Unit Average Relative Saturation Initial #(-m)

TCw 0.67 (*23) 194PTn 0.61 (*15) 64TSw-1 0.65 (*19) 2ATSw2-3 0.65 (*19) 232CHnv 0.90 38

The steady-state vertical pressure distribution from the approximate

center of the model domain in the Case 1 simulation was used as the initial

pressure distribution for the Case 2 simulation to ensure that the specified

boundary conditions were consistent with the internal solution. In the

Case 2 simulation, the shafts, drifts, and MPZ subdomain elements were included

in the model domain. The steady-state pressure distribution from the center

of the model domain in the Case 1 simulation-was also used as the initial

pressure distribution for the transient simulation.

2.4 BOUNDARY CONDITIONS

The surface boundary condition for the simulations was specified as a

flux boundary with fluxes representing recharge rates of 0 or 4.0 mm/yr.

This 4.0 mm/yr recharge rate is comparable to the 4.5 mm/yr recharge estimate

cited by Montazer and Wilson (1984). This estimate assumes that after evapora-

tive losses, 3% of the average annual -precipitation (# 150 mm/yr) at Yucca

Mountain percolates back into the groundwater flow system. This 4.0 mm/yr

recharge flux was alternated with a zero-flux surface boundary condition for

the transient simulation to approximate the seasonal distribution of precipita-

tion. The lower boundary was specified as a fixed pressure (atmospheric)

boundary,--representing the water table.

The regional dip of the units in the vicinity of the ESF at Yucca Mountain

is approximately 6 degrees to the East (Scott and Bonk, 1984). This creates

a driving force in the lateral (down dip) direction which is approximately

10% of gravity. This was not directly accounted for in the model. The left

or West boundary of the model domain, however, was specified as a zero-flux

or "no-flow" boundary so that water could not move up dip across the Ghost

2.16

Dance Fault. The North, South, and East boundaries of the model domain were

specified as fixed-pressure boundaries, with pressures corresponding-to the

initial conditions previously described. The East boundary was specified as--

a fixed-pressure rather than a zero-flux boundary to create a lateral flow

component approximating the effects of the regional dip of the stratigraphy.

The North and South boundary specifications reflect our uncertaMnty ii the \

influences of topography and structural features within the area represented

by the model in controlling preferential local recharge. However, the

combination of the specified boundary conditions creates a preferential recharge

area corresponding to the locations of Coyote Wash and the Ghost Dance Fault,

which is consistent with our conceptual model. Future simulations will utilize

a modified version of PORFLO-3 which incorporates a gravity tensor so that

the effects of the lateral driving force created by tilted beds are directly

accounted for.

2.17

3.0 COMPUTER CODE DESCRIPTION

The flow simulations were performed with the PORFLO-3 computer code

(Runchal and Sagar 1989; Sagar and Runchal 1989). This code is designed to

simulate fluid flow, heat transfer, and mass transport in variably saturated

porous and fractured media. Three-dimensional simulations in eithersteady_

or transient mode can be performed. The steady-state solution can either be

obtained directly by putting the time-dependent term to zero in the governing,,,

equation or by solving a transient problem until the solution becomes steady.

The first option is very efficient and is for the steady-state simulations

reported here.

Fractures, faults, shafts, and drifts can be-discretized as full three-

dimensional features or embedded as one- and two-dimensional features. The

basic assumption in this formulation in which one- and two- dimensional features

are included in three-dimensional calculational domains is that the pressure

values at grid nodes represent both the subdomain features within calculational

cells (ie. shafts or faults) and the calculational cells (ie. adjoining rock).

Thus there is no transfer of fluid between the cells and the subdomains within

the cells. Such a transfer occurs only across cell boundaries. As a result,

calculated pressure values are not as accurate close to the subdomain features

as they would be if the features were discretized as a full three-dimensional

features. Embedding one- and two-dimensional features within a three-

dimensional domain is much more efficient, however, because it requires fewer

nodes and therefore less computational time. Because of the large calculational

domain used for the simulations reported here, the shafts, drifts and Ghost

Dance Fault are included as one- and two-dimensional features rather than

being discretized as full three-dimensional features.

The options available in the code for describing hydraulic properties

of porous media are the Brooks and Corey (1966), or van Genuchten (1978) water

retention models with hydraulic conductivities determined by utilizing either

the Mualem (1976) or the Burdine (1953) formulae. An additional option is to

provide the water retention and the relative hydraulic conductivity curves in

tabular form. This last option is useful for specifying the characteristic

curves for fractured media that are represented as equivalent porous media.

3.1

If tabular values are input for the hydraulic properties, the code linearly

interpolates between the values as needed while solving the unsaturated flow

equation.

The PORFLO-3 code uses the pressure based formulation of the flow equation

which is discretized using an integrated finite difference approach. Alternate

methods of solution are provided to solve the set of algebrikc/equati"ns. X

iterative conjugate gradient solution method (Kincaid et al. 1982) is employed

for the results reported here.

3.2

4.0 RESULTS AND DISCUSSION

The following results represent steady-state and transient simulations

with upper-bound MPZ properties (see Table 4.4), at a recharge rate of

4.0 mm/yr. The 4.0 mm/yr recharge rate is comparable to the 4.5 mm/yr recharge

estimate cited by Montazer and Wilson (1984). Other net in itratiobnestm ates

for Yucca Mountain range from 0.1 to 0.5 mm/yr (Peters et al. 1986). Therefore,

the results of these simulations should be considered as conservative estimates

of the effects of the exploratory shafts on the hydrology of the Yucca Mountain

site in the vicinity of the ESF.

4.1 STEADY-STATE SIMULATIONS

Direct steady-state solutions did not converge at a recharge flux of

4 mm/yr without modifying the hydrologic properties of the Tiva Canyon welded

tuff unit (TCw). The primary cause of the divergence is the abrupt change in

composite hydraulic conductivity of unit TCw as shown in Figure 2.5a. With

high recharge fluxes, pressures develop that indicate the beginning of fracture

flow (i.e., saturation becomes close to 1). However, the recharge is not large

enough to sustain this fracture flow. The result is that the solution

alternates from matrix to fracture flow between iterations and will not

converge. Scaling up the composite saturated hydraulic conductivity of unit

TCw by a factor of 25 for the Case 1 simulation was sufficient for successful

convergence of the direct steady-state solution.

By scaling up the saturated. hydratulic conductivity of unit TCw for the

Case 1 simulation, higher recharge fluxes could pass through unit TCw and

into the underlying unit PTn without jumping from fracture to matrix flow

between iterations. It is not clear if the solution obtained with this

modification-is a realistic one because fracture flow in unit TCw was

effectively-eliminated. However, the Case 1 simulation was only used to

establish initial and boundary conditions consistent with the internal solution

at a recharge rate of 4 mm/yr for the Case 2 simulation, so the absence of pulse

infiltration resulting from fracture flow was considered to be relatively

insignificant.

The saturated hydraulic conductivities measured on core samples of unit

TCw range over 4 orders of magnitude (Peters et al. 1984). In addition to this

4.1

variability, the hydrologic properties of the fractures reported by Peters et

al. (1986) are estimated rather than measured directly (Wang and Narasimhan

1985). Therefore, scaling up the composite hydraulic conductivities of unit----

TCw by a factor of 25 should still yield hydraulic conductivities within the

range of uncertainty of the equivalent continuum approximations. ,Without

this hydraulic conductivity modification, direct steady-state solutiows at '

high recharge fluxes would require much finer spatial discretization inK'the

vertical direction. Another alternative would be to run a transient problem</-

until steady-state is reached. Numerical stability could then be achieved by

reducing the size of the time steps. Both of these alternatives, however,

are much more expensive computationally.

A steady-state relative saturation profile through the center of the

model domain from the Case 1 simulation is shown in Figure 4.1a. The shafts,

drifts, MPZs and Ghost Dance Fault were excluded from this simulation. If one

could assume that the initial conditions specified for the Case 1 simulation

represent a quasi-steady state for the systenm-(in the absence of the ESF),

and that the hydrologic properties are reasonable, then-the higher relative

saturations obtained in this solution suggest that-the long-term infiltration

rate at Yucca Mountain is considerably lower than the 4 mm/yr assumed here.

The steady-state pressure distribution corresponding to the relative saturation

profile of Figure 4.1a -is shown in Figure 4.1b. This pressure distribution was

used for the initial and fixed-pressure boundary conditions specified in the

Case 2 simulation.

As with the Case 1 simulation, the direct steady-state solution for the

Case 2 simulation did not converge at a recharge flux of 4.0 mm/yr without

modifying the composite saturated hydraulic conductivity of unit TCw. Scaling

up the composite saturated hydraulic conductivity of unit TCw by a factor of

15 was sufficient for successful convergence in the Case 2 simulation. A

smaller hydraulic conductivity scaling factor could be used for the Case 2

simulation because the initial and boundary conditions for this simulation

were more consistent with the internal solution than the uniform average

saturations in each unit that were imposed as boundary conditions in the

Case I simulation.

4.2

t

a.0,

10Q.0

*SIILO

-ii0.0 -

0.0~

*50.0*

.1W.Q .

is 2S.0 -

P.U' 1.O

zo -O.e-

6-

N W~Li

*450.0 *no.0.

i_

'550.B.. _ _ _a., 0.2 0.3 0.4 5.1 0.a

SnTuRAT IONO.? i.' 0.5 1.5

"s5n.n . . . . *

-awou -'521 '5W.0 60-.0 -44me -.7. 2 30.0 -RSO 0.0 -2A.8 100IL.0 -21.0 .

PRESSURC MI1

FIGURE 4.1a. Steady-State Relative SaturationProfile from the Center of the ModelDomain for the Case 1 Simulation(without Shafts and MPZs)

FIGURE 4.1b. Steady-State Pressuro/Distribution fromthe Center of the ModelDomai-n for theCase 1 Simulation,(Initial ConditionsSpecified for Case 2 imulation)

., ,. I

A steady-state relative saturation profile through the center of the model

domain for the Case 2 simulation is shown in Figure 4.2a. The shafts, drifts,

MPZs, and Ghost Dance Fault were included in the Case 2 simulation. Coomparison

of Figure 4.2a with 4.1a indicates slightly higher relative saturations in the

welded units for the Case 2 simulation with very little change in the relative

saturations of the non-welded units relative to the Case 1 simulation. The

differences in the relative saturation profiles shown in FiguIres 4.1I a Vd 422a

show the influence of the imposed boundary conditions on the internal solution.;

A steady-state pressure distribution through the center of the model

-domain for the Case 2 simulation is shown in Figure 4.2b. This pressure

distribution corresponds to the relative saturation profile depicted in Figure

4.2b. This figure is also provided for comparison with Figure 4.1b which showsthe pressure distribution from the Case 1 simulation at the same location.

A steady-state relative saturation profile and pressure distribution

through the line elements representing ES-1 for the Case 2 simulation are shown

in Figures 4.3a and 4.3b, respectively. The slight pressure/saturation

discontinuity at the interface between unit-TSwl and-TSw2-3 reflects slight

differences in the properties of these units. The maximum relative saturations

obtained from the Case 2 simulation for the welded units TSw1 and TSw2-3 are

0.957 and 0.936, respectively. The hydraulic conductivity and relative

saturation curves shown i-n-Figure 2.7 indicate that water in the fractures is

essentially immobile at relative saturations *of less than about 0.98.

Therefore, the results of this simulation suggests that fracture flow, as

represented with this equivalent-continuum approximation, will not occur in

these welded tuff units.

In one-dimensional simulations of vertical flow through the unsaturated

zone at Yucca Mountain using the TOSPAC code, Dudley et al. (1988) used a 2303

node model extending from the water table (0.0 m) to the surface of Yucca

Mountain (530.4 m). Results of these one-dimensional steady-state simulations

show the TSw units to be saturated or near-saturated, and thus amenable to

fracture flow, at recharge rates greater than 1.0 mm/yr. This is contrary to

the results of the three-dimensional PORFLO-3 simulations reported here.

This difference is attributed to the nature of the problem and the boundary

conditions specified for the three-dimensional simulations which allow laterally

4.4

I

0.0

-50.0-

0.0'

-50.0 *

-100.0 '

zCI.-

0a.N

-r50.0

n,Ix.l0 -

zo -zsn.a

U,0

-RI-".aIl

*2,SO.0 *

-100.0'

-150.0 '-150.0-

-s5n.0

a.' 0.2 5.3 0.1 0.S d.SSnTURAT ION

0.1 0.3 i.3 D .0-150.04 ! .I 0 ,

-sso.#. -00.0 -4o.0 -30.0 -3m.0 -250.6 -m. -150.0 -100.0 *50.0 0.PRCSSURE 1 l)

a

FIGURE 4.2a. Steady-State Relative SaturationProfile from the Center of the ModelDomain for the Case 2 Simulation(with Shafts and MPZs)

FIGURE 4.2b. Steady-State Pressur Distribution fromthe Center of the Model domain for theCase 2 Simulation,(wit' Shafts andDrifts) /

'V /~~~~~~~~~~~~~~~~~~~~~~~~~~~~

I

0.0* 0.0

.50.0!

-110.0 -ISO. 0

-*2.0,

0

R

-25.020) -l10.0

vi

*4W7.0l

/-110.0.-110.0 -

0.7 0.5 S.9 I.00.0 I.1 0.2 0.3 o.I 0.S 0.5SnTURnT ION

4W3.-IWO.S -qsb.0 -qo 310.0 *3W.0 -A1.0 Vg .2w.0 -gso.q* 0 .90'.0PRESSUREC lIIi /

0.0

FIGURE 4.3a. Steady-State Relative Saturation Pro-file from the Center of the ModelDomain for the Case 2 Simulation

FIGURE 4.3b. Steady-State Pressurae Distributionthrough Line Elements RepresentingES-1 for the Casea2 Sjm/ulation

/ /!

flowing water to exit the model domain. Any lateral diversion of vertical

fluxes above the repository horizon decreases the likelihood of the repository

horizon reaching saturated or near-saturated conditions, unless the laterally

diverted water encounters a preferential flow path such as an MPZ around a

shaft, or a fault zone.

A vertical cross-section of Darcy velocity vectors through the~plane >

that intersects ES-1 for the Case 2 simulation is shown in F l re 4.4. Slig'ht

perturbations in the flow field caused by ES-1 and the MPZ around it are evident

at the interface between units TSwl and TSw2-3 and jitst'above the drifts

connecting the shafts in the repository horizon at a depth of about 275 m. A

potential for significant lateral flow above the interface between units PTn

and TSw1 is also indicated in Figure 4.4. This lateral flow is consistent

with the results of Rulon et al. (1986) and Wang and Narasimhan (1988) who

also show a potential for significant lateral flow above the repository horizon

using two-dimensional cross-sectional flow models.

A horizontal cross-section of Darcy-velocity vectors through the lower

part of unit TCw, just above the interface with unit PTn from the Case 1

simulation is shown in Figure 4.5. A horizontal cross-section of the same

plane depicted in Figure 4.56 is shown in Figure 4.6 for the Case 2 simulation.

The effects of the shafts are clearly evident from the perturbations in the

flow field shown in Figure 4.6, relative to Figure 4.5.

The steady-state pressure distribution corresponding to the flow field

depicted in Figure 4.6 for the Case 2 simulation is shown in Figure 4.7.

These pressure contours also indicate slight perturbations in the flow field

resulting from the presence of the exploratory shafts. The relative saturations,

corresponding to the pressure distribution shown in Figure 4.7 are shown in

Figure 4.8.

A horizontal cross-section of Darcy velocity vectors through the upper

part of unit TSw2-3 just below the interface with unit TSwl from the Case 1

simulation is shown in Figure 4.9. A horizontal cross-section of the same

plane depicted in Figure 4.9 is shown in Figure 4.10 for the Case 2 simulation.

Again, the effects of the shafts are clearly evident from the perturbations

in the flow field shown in Figure 4.10, relative to Figure 4.9.

4.7

1 107891712

CASE 2 - WITH SHAFTS AND MPZSSCALED VELOCITIES, TRUE DIRECTIONS

;4-!1- -g S-- - : - - ---:l-- -1- -lli -iill -o VYYYY Y Y.Y V Y Y V Y V VVVYYVVYVY Y

N -i- |-l---- ---- |--------|-----|-----||+-----|f4--; v v vv vrv v v v vv v v v

7 v vv v r v v v v v v v v v eve v v vvs v v v '

VY VY Y Y Y v Y Y Y Y v v Y V Y Y Y Y Y Y

v Y YY Y Y Y Y Y Y Y v Y Y YYYY v Y YV' v Y" Y v '

i- - I . i . .- nuj -- .-. -.f. .1

vvvvv ~~~~v v v vv v v vvVVVVV vY V v

0.0 50. 0 100.0 150. 0 200. 0 250O. 0 30O.0 35'0.0 400.0 45D.0 500.0 S50.0a 600.0 650.0

SCALE rACTOR O.SO X TIME - I.OOE30

FIGURE 4.4. Vertical Cross-Section of Darcy Velocity Vectors through Plane Inters cting

Plane Intersecting ES-1 for the Case 2 Simulation

CASE 1 - NO SH-RFTS OR MIPZ5SCALED VELOCITIES, TRUE DIRECTIONS

A AA A A 4 A I I I I All& A A ill I I I /

A IA A A A A A A A I AI A I A AA a I A I a ' '

0AAII I A A A AI I A A IIII A A I LLJ V

~A AA A I A A AI A I A A I AAAAALJ41 IVV

I, - - -~~Am

Y !? V I V V V tY .' V~.

v vyvT V T I Y !1I V V !IV.

I TTI Y V V V IV V V I T T VI I I

I t $ 4 $ i I ~~~~I 4 41i14Ci. ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ U IL

\

0.0 S0.0x TTIlE -I.ME

=.O 95L

FIGURE 4.5. Horizontal Cross-Section of Darcy Velocity Vectorsthrough the Lower Part of Unit TCw for the Case 1Simulation __ -

.- I"

CRSE 2 - WITH SHRFTS AND MIPZSSCF.LED VELOCITIES, TRUE DIRECTIONS

U .4A AA A A A A I AAAI AI III A lAAAI I A

~~AY'Ak A A A ~~A A A A IA AAAII IIA &AlA A

~A k A A A1-, ~ A IIA AA AI i4 4 I 'f

o A~~~~~~~~ '. ~ ~ ~-I I 'A

V V

V I I V II I I I I-

IAV V I T IV 1V!VT TV I IV I1 I

v Y1 I TI 4 I I~IIVTITVI iI It t

0.0 5C.3 1M.0 1W.U

scpLc rAcTUR 0.50x TH~E IEM

sn.W A.

FIGURE 4.6. Horizontal Cross-Section of Darcy Velocity Vectorsthrough the Lower Part of Unit TCw for the Case 2Simulation

4.9

0

FIGURE 4.7. Steady-State Pressure Distribution in Horizontal Cross-Section throughLower Part of Unit TCw for the Case 2 Simulation (corresponding to''Figure 4.6)

I

I-.

FIGURE 4.8. Steady-State Relative Saturations in Horizontal Cross-Section throughLower Part of Unit TCw for the Case 2 Simulation (corresponding io

Figure 4.7) /

9- ~

C!

CASE I - NO SHAFTS OR MPZSSCALED VELOCITIES, TRUE DIRECTIONS

'PITYT-7 -- --TiTfTTTOTT.T -- 4

jail a A £ £ A~~~~ A Al4 AIII 4441 i

AIA A A I A & A AAAA AiIft11 f4*4e fle V,AAA A AI A A A A A A A I I A111~4444vf'

LAAA A A A £ A A A A _ AF 4 1e4 4' 44' w -

IlY II I -.ZI !X jv Iv I 1% 1k 'k1IY1 I I I I I I 1V1~4

VI4 4I 4 V V I I I IIItIIi~II 'k %

IV I I_ __ _ I I

N.

0._ _ . . .

[ T I

0.2 50.0 1M.2 15_0.

50VU. MTORc 0.502~. 0 250.0 X~.O 350.0 46C.0 g~f.GS LO~ 5jf0

X TIM1 -1.M0

.

6 Gi 10.0

FIGURE 4.9. Horizontal Cross-Section of Darcy VelocityVectors through the-Upper Part of Unit TSx2-3for the Case 1 Simulation-

~~~ ~~I I W M L J I 2

CASE 2 - WITH SHAFTS RNO MPZSSCMLEO VELOCITIES, TRUE DIRECTIONS

a

R il) ;l i a a a A i A i AAAAAA A JAA IA A 4 4 :

o Lit I I A A A I I A A AI A A A A AI Ii W e

e A4~ A A A ^ k k 1 ^ ^ | ^ ^ s ^A A A I A J A A1 4 W444 X f W -_A 1 %&@s1 A A A A A IA. 4. A A 4 1 4 4 4A.A I I *If

1+\ ^ A* I 4 A 4_ _s . A A 1 ti- 4 : : ; I

o vs 1 . ,1 1 * 1 V 1 b. -s 5 s se

8- srt * 1 1 V 1 V I I 1 1. b. I IIII It _ _,

_ 1~s'I V 1 1 1 v 1 1 v 1¶111IVI I A4 ' 1* IV,! 1 1 V V I I V ~~~~~I I I I I II 1 V 1 I iVI I

e [s t s 1 1 I 1 1 V I Y T T I I I V '. 4 1 1 .s1 s

. 11ft1 1t I t V 1 1 1 1 t 1 t l tl l 1 tY V I 1 1 '

0.0. SD.O ItlD.O I50.0D 20.0 2 .0.0 350.0 4qM 1!OO 5C0. 0 Ssl.tl C .0 D S.0

SC rFRCIORC 0.50 X TIME -Iz.=

FIGURE 4.10. Horizontal Cross-Section of Darcy VelocityVectors through the Upper Part of Unit TSx2-3for the Case 2 Simulation

4.12

The steady-state pressure distribution corresponding to the flow field

depicted in Figure 4.10 for the Case 2 simulation is shown in Figure 4.11.

Again, as shown in Figure 4.7, these pressure contours also indicate-slight

perturbations in the flow field resulting from the presence of the exploratoriy

shafts. The relative saturations corresponding to the pressure distribution

shown in Figure 4.11 are shown in Figure 4.12. As shown by the previous

figures, the boundary conditions, fault placement, and symmetry of this modNi

create a preferential recharge area on the left or West side of the model-,

domain, which corresponds to the area where Coyote Wasfi-crosses the Ghost

Dance Fault.

A horizontal cross-section of Darcy velocity vectors through the repository

horizon at the 310 m depth in unit TSw2-3 for the Case 1 simulation is shown

in Figure 4.13. A horizontal cross-section of the same plane depicted in

Figure 4.13 is shown in Figure 4.14 for the Case 2 simulation. The effects

of the shafts are not as evident at this depth as they are near the interface

between units TSw1 and TSw2-3 and in unit TCw.

The steady-state pressure distribution corresponding to the flow field

depicted in Figure 4.14 is shown in Figure 4.15. The relative saturations

corresponding to the pressure distribution shown in Figure 4.15 is shown in

Figure 4.16. As shown by the flow fields in the previous figures, this pressure

distribution also indicates that the effects of the shafts are less significant

at the depth of the proposed repository than at shallower depths.

4.2 TRANSIENT SIMULATIONS

An objective of this study was to determine the potential effects of the

ESF on the moisture distribution at Yucca Mountain, with emphasis on short-term

characterization work. Therefore, a transient simulation was also performed,

using a-time-varying recharge rate. The surface boundary flux for this

transient simulation was represented with a step function so that a recharge

rate corresponding to 4 mm/yr in 4 months (the rainy season) alternated with

no recharge for 8 months. This seasonal recharge was maintained for a 20

year simulation period. This was considered to be an adequate time frame in

which to-conduct subsurface characterization work at Yucca Mountain. No

modifications to the saturated hydraulic conductivity of unit TCw were

4.13

FIGURE 4.11. Steady-State Pressure Distribution in Horizontal Cross-Section through

Upper Part of Unit TSw2-3 for the Case 2 Simulation /

I

CASE 2 - WITH SHAFTS AND MPZSTI lET-

PonrLO301107891712

1 CONTOIA 2.031I 121909 1700

P-.u.'

xSe.0 gm /E *'0.0

T IMCt - I .W030

FIGURE 4.12. Steady-State Relative Saturation in Horizontal Cross-Section throughUpper Part of Unit TSw2-3 for the Case 2 Simulation (corresponding toFigure 4.11)

CASE 1 - NO SHAFTS OR MPZSSCALEO VELOCITIES, TRUE DIRECTIONS

32- 5t

44 j4

AL~~t A £ A A A A £ ~~~~A A A £Ai i ld14 .

hA I A A A A A A A I 14 444'~

IAAA A A A A A A A I A 4 444 4444.'. V v,AL JA AI A 'A A f fr r f.'r

V''V V I V V ' 'I ' V I IVIII V 'V I V I ~ ~v 'V I 'V IV I t A 1 k '

~~ii$ I $ $ 4 $ I $ I $ I Itt~~~~lt~tlI AV . N

IN

ToO. ....

0.0 SO. OSC9LE FRC-O 0.5i 3 ~~~~~~~~~x TIM -1.~00

m.: a3.1

FIGURE 4.13. Horizontal Cross-Section of Darcy VelocityVectors through the Proposed Repository Horizonin Unit TSw2-3 for the Case -Simulation

,.w5....

CASE 2 - WITH SHAFTS AND MPZ5SCHiLEO VELOCITIES, TRUE DIRECTIONS

C!

4 -~~~ ~~~ 4 ~~~ 4 4~~ I ALA A A A h i £I A .

AA1A A A A A A' A A A I IAA.II A AAAA44 W .'

AAAAA AI A A IIA AA A A A A IJJ4II44 -F *

~LAAA A A A A . I IA A A A1111I4j114444.'

A LA A A I. A A A ~~~~~~ A A I> A I,(IffIf 4144 4.e

v A Y v I I I~ ~~ ~~ ~~~~~~~~~~~ I v L A 4 4 '.r%:7V IA IV I 'V I I I 'V I V I 1 ' 1% 4 * 44~ . - - -

1V 1, I I V I IV V v I V 1 I A AA

Blilil V 'V I I V V~~~~ I I A I11 AI A4&

IlyVI i V V ~~~~~v IV ' I 1I 'I I I ItI I I

~~I~tVI~ I II Vt I V I V TIY I II IVV''' I

fifl 100 1.0 110.0 MJ0.0 250.0 3M.0 M0. 4~0_ qSLA SC6.0 -SA60.

SCF%.r- rRCTOR C.50 x TIM - 1.

FIGURE 4.14. Horizontal Cross-Section of Darcy VelocityVectors through the Proposed Repository Horizonin Unit TSw2-3 for the Case 2 Simulation

4.16

CASE 2 - WITH SHAFTS RNO MPZSP -

POyrn-o301 107891712

1 CONTOUR 2.03)11219891711 I

FIGURE 4.15. Steady-State Pressure Distribution in Horizontal Cross-Section throughthe Proposed Repository Horizon in Unit TSw2-3 for the Case 2 Simulation(corresponding to Figure 4.14) -

CASE 2 - WITH SHAFTS AND MPZSTHET-

PoRrL0301107891712ICONT0OJ 2.031I 1219091605 I

9

P.

9C.

9M.

9

N

9M.

9IC

961-

-- 5 - a 0.86 0.86

- 0,87 0.87

0 .88 .- I N 0

0. 90

O'.90

I

/ 0.90,_--� 0.90 -IIII

00.919

9K

aSt

9A

9a

------------ 89 I

:,Be.0.07 -

n-An -

I~~~~~~~~~~~~~~~~~U ole

6.0 Ii.0 IdO.0 1�0.0 K;3.0 r;3.0 3dO.O 3�0.0 460.0x

qs�.b S60.0 , , 55b.0 5do.0 '. 6�0.0I ., / T I "E - I .OM,30 -

FIGURE 4.16. Steady-State Relative Saturations in Horizonal Cross-Section throughthe Proposed Repository Horizon in Unit TSw2-3 for the Case 2Simulation (corresponding to Figure 4.15) / 'I,

necessary for the transient simulation because the size of the time steps

were reduced to achieve numerical stability.

During this simulation period there were no appreciable changes in the ---

relative saturations of any of the units other than unit TCw. ,The relative

saturation profile for unit TCw from the center of the model Aoman fo the

transient simulation is shown in Figure 4.17. The relative<~ urat pro j

for unit TCw along the plane representing the Ghost Dance Fault for the

transient simulation is shown in Figure 4.18. The re, tive saturation profi

for unit TCw at the location of ES-I for the transit/nt.imulation is shown in

Figure 4.19. The plot of time zero in Figures 4,47 4 18 a 4.19 represent

the initial condition from the steady-state solutiofrom the center of the

model domain in the Case 1 simulation.

Comparisons of Figures 4.17, 4.18 , and 4.19 indicate that with the

material properties used for these simulations, the MPZs around the shafts

create a preferential flow path for/,yater percolating down from the surface.

The Ghost Dance Fault also acts asia preferefti-flow path, but its effects

are not as significant with the particul prpperties ehosen to represent it

relative to the MPZs and shafts. At the relative saturations shown in these

figures, fracture flow, as represented with an equivalent continuum

approximation, is occurring in unit TCw. \

Simulations by Wang and Narasimhan (198'6), using different constant and

pulse infiltration rates applied-to fractures in unit TCw, indicate that most

pulse effects are eftectively daimped'eut by units TCw and PTn before water

infiltrates down into unit TSwl. Ctlculations by Travis et al. (1984) indicate

that the penetration'distance of pulses of water into fractures contained in

densely welded tuffs similar to-unit TCw is of the order of 10 m or less if

the frap~ure-apertures are 100-micrometers or less. Therefore, although

frac re may\be significant in unit TCw, these simulations suggest that

epi'odi'c pulses\of water from natural precipitation should not propagate via

fracture flow into the underlying units due to capillary imbibition into the

rock'"matri x

These-,transient simulations suggest that there will be no appreciable

change in the"-moisture content distribution below unit TCw within 20 years

4.19

K'-I ..

Relative Saturation Profiles for Unit TCw from thethe Model Domain for the Transient Simulations

A/

I

/ N.

IIN

I,

I/ /

0.0-

-5.0-

'I

L

,)

I.-

in

N

-I0.0-

-I5.0-

1'l,

/ i

<I

j/

~I/ /II

0a

x

V

0

I~~~

I:A.

0.g~d ,.:

LEGENOTHIC0.000

120.0365.31581.1826.3107.3653.5234.5479.7060.7305.i

i /.7%-n ,

o.52 0.81RELRTIVCE SnTuRAT ION

o:9s

FIGURE 4.18.

/I~~~~~~N

Relative Saturation Profiles for Unit TCw along the Plane Representin' tGhost Dance Fault for the Transient Simulations

A~~/

he

,/

i

"13

Relative Saturation Profiles through Line Elementsin Unit TCw for the Transient Simulations

.,

after the construction of the ESF. These simulations, however, did not account

for possible changes in moisture content resulting from additions ofdrilling

fluid into the system during the excavation process. Work by Buseheck and.

Nitao (1987), however, suggests that imbibition of the drillingI fluld used

during the drill-muck-mining excavation process will be minimal. Thefore,

even though shaft construction will alter the fracture dens t's an

permeabilities of the tuff units, the ambient moisture contents and >

distributions in the vicinity of the shafts below TC s ould not change

significantly within a reasonable time frame for subsurface characterization

work. /

The MPZs around the shafts in unit TCw a pear 2 ,eate a preferential flow

path for water when fracture flow occurs in this-unit. This fracture flow

appears to be damped with depth, however, so that no fracture flow occurs in

the underlying units. However, it must be emphasized that these conclusions

are specific to the initial and boundart conditions and material properties

used for these simulations and do/not the possibility of lateral

flow intersecting preferential flow pat? i resuft~ing in locally saturated

conditions and fracture flow in the lower units' -

/ - /

K~~~~~~~~~~~~~~~~~,

* K~~~~~~~~~~~~~N ~~~~~~~~,/~ W

/ ~ ~K ~ .

4.23

5.0 CONCLUSIONS AND RECOMMENDATIONS

In this model, the exploratory shafts, drifts and the Ghost Dance':Fault..

are embedded as one- and two- dimensional features within three-dimensional

calculational elements. These features are assumed to have tMhir own".distinct

properties, but the results obtained from this model are no as ccur e

close proximity to these features as they would be if the fea ures were~dis- \

cretized as three-dimensional elements. Considering the lack of data regarding

properties of the Ghost Dance Fault and MPZs, and the co paratively large

calculational domain of the model, the results presented here should be reason-

able. The size of the calculational domain used/inithis, ?nal~sis is considered

necessary in order to include the Ghost Dance Fault and' the exploratory shafts

while attempting to minimize boundary effects. If aInore accurate representa-

tion of specific small-scale features is desired, the model domain should be

reduced to a smaller scale and computational cell sizes reduced accordingly.

Larger scale approximate solutions,, sucr-asthose presented here, can readily

be used as input to more finely discret4zed smaller-scale models within the

domain of the larger model. On a sufficiently smaller/'scale, the shafts may

be included as three-dimensional features. '

Several important conclusions can be'\drawn from these simulation results.

The simulations suggest thatNfracture flow,\'as, approximated by the equivalent

continuum model (i.e., when hydrologic properties are represented as shown in

Figures 2.6 to 2.8), does not occur at the level of the proposed repository

(i.e., in the Topopah"Springs/we1ded unit) at recharge rates up to 4 mm/yr.

Discrete fractures or faults' were noit'simulated, however. Therefore, these

results do not preclude-the possibility of preferential flow reaching the

proposed repository horizor-Nthrough regions of major structural features.

The Ghosti iae"-Fault is included in this model, but the hydraulic properties

used/to describe the fault zone are estimates due to a lack of data. These

simulations also suggest that there will be no appreciable change in themonistureicontent distribution below unit TCw within 20 years after the

construction'of the ESF. This time frame was considered to be adequate for

subsurface characterization work.

5.1

Previous one-dimensional simulations have indicated that the repository

horizon will be saturated or nearly saturated and thus amenable to fracture

flow at recharge rates of 1.0 mm/yr. This is contrary to the results of the----

three-dimensional simulations reported here. This difference cpn be primarily

attributed to the dimensionality of the problem and the nature of.thethree-

dimensional flow field such that water can exit from the mod v omal thro g

lateral boundaries. The results of these three-dimensional simulations pport

previous two-dimensional modeling studies that show a potential for signifkant

lateral flow above the proposed repository horizon. ALateral diversion of

vertical fluxes above the repository horizon dec gase the/likelihood of the

repository horizon reaching saturated or near-saturated condftions unless the

laterally diverted water encounters a preferential flow path or structural

feature such as a fault zone.

Post-closure thermal effects and vapor-phase transport of moisture caused

by heat generated from the repository-were not considered in this study.

However, these factors may have a s'ignificiftt-affect on the performance of

the repository. Therefore, future postZcTosure modeling studies should consider

non-isothermal flow and vapor-phase transport ofuidoisture. The exploratory

shafts and MPZs may become more important for'future smaller-scale model

simulations when coupled fluid flow, heat\.transfer, and mass transport are

considered.

It will be impossible to characterize all of the subtle heterogeneities

in the properties/of~the tuff units at-Yucca Mountain in sufficient detail to

predict all possible outcomes associated with a particular characterization

activity or the performancexof the repository after waste emplacement. However,

at a minimum, future simu1atio'hs with intermediate and larger scale multi-

dimensi oi-mo~Iels of the proposed repository at Yucca Mountain should consider

the topography o1\the site and surface expressions of major fractures and

faults/to determine potential areas for preferential local recharge. Then,

spatially non-uniform, time-varying recharge fluxes should be applied as surface

boundary conditions to obtain the most accurate representations of flow fields

in the naturaT-system at depth. In addition, multi-dimensional stochastic

modeling approaches should be used to evaluate the effects of the variability

5.2

and uncertainty of the hydrologic properties of the tuff units on calculationsof repository performance objective criteria such as groundwater travel time.

/\ \

\/ \* A.. K I

I

/1

1"

5.3

_ rS

6.0 REFERENCES

Brooks, R. H., and A. T. Corey. 1966. 'Properties of Porous Media Affiecting-Fluid Flow," J. of the Irrig. and Drainage Div., Proceedings of the Amer.Soc. of Civil Engrs., IR2, pp. 61-68. / X

Burdine, N. T. 1953. "Relative Permeability Calculations from PoreNSizzeDistribution Data," Petroleum Transactions, AIME, T. P. 3514,Vol 19 pp .71-78.

Buscheck, T. A., and J. J. Nitao. 1987. Estimates of"tpe Hydrolooic Impactof Drillina Water on Core Samples Taken from Partially Saturated Densely Welded-Tuff. UCID-21294, Lawrence Livermore National Laboratory,,Livermore, CA.

Case, J. B., and P. C. Kelsall. 1987. Modification of RMckMass Permeabilityin the Zone Surrounding a Shaft in Fractured, Welded Tuff. SAND86-7001, SandiaNational Laboratories, Albuquerque, NM.

Dudley, A. L., R. R. Peters, J. H. Gauthier, M. L. Wilson, M'.S. Tierney, andE. A. Klavetter. 1988. Total System Performance Assessment Code (TOSPAC)Volume 1: Physical and Mathematical Bases. SAND85-0002, Sandia NationalLaboratories, Albuquerque, NM.

Fernandez, J. A., T. E. Hinkebein, and 3h B. Case.---1988. Analyses to Evaluatethe Effect of the Exploratory Shaft on Repository PerfYmance of YuccaMountain. SAND85-0598, Sandia National Laboratories, Albuquerque, NM.

GE/CALMA. 1986. Interactive Graphic and Information Systems, Product Numbers:CALO114, CAL0118, CALO11, CALOO94, CALOO82, CAL0107 to CAL0109, Department6310, Sandia National Laboratories, Albuquerque, New Mexico.

Kincaid, D. R., J. R. Res-pess, D. A. Young, and R. G. Grimes. 1982."ALGORITHM 586 ITPACK 2C: A FORTRAN Package for Solving Large Sparse LinearSystems by Adaptive Accelerated Iterative Methods," ACM Transactions onMathematical Software,\8(3), pp-.-302-122.

Klavetter, E. A., and R. R',Peters. 1986. Fluid Flow in a Fractured RockMass. SAND85-0855, Sandia National Laboratories, Albuquerque, NM.

Montazer.-P- and W. E. Wilson. 1984. Conceptual Hydrologic Model of Flowin the-Unsaturated Zone, Yucca Mountain, Nevada. Water-ResourcesInvestigations Report 84-4345, U.S. Geological Survey, Denver, CO.

Mialem, Y. 1976. "A New Model for Predicting the Hydraulic Conductivity ofUnsaturated Porous Materials," Water Resources Research 12(3), pp. 513-522.

Nitao, -J. J. -1988. Numerical Modeling of the Thermal and HydrolooicalEnvironment Around a Nuclear Waste Package Using The Equivalent ContinuumApproximation: Horizontal Emplacement. UCID-21444, Lawrence Livermore NationalLaboratory.

6.1

PBQ&D. 1985. Parsons, Brinckerhoff, Quade, and Douglas, Drawing No. R06940,GE/CALMA No. 0119.

Peters, R. R., J. H. Gauthier, A. L. Dudley. 1986. The Effect of 'Percolation--Rate on Water-Travel Time in Deep, Partially Saturated Zones. SAND85-0854,Sandia National Laboratories, Albuquerque, NM.

Peters, R. R., E. A. Klavetter, I. J. Hall, S. C. Blair, P.,A. Be t r, a .Gee. 1984. Fracture and Matrix Characteristics of Tuffacebus/4aterials fr'DmYucca Mountain, Nve County. Nevada. SAND84-1471, Sandia National Laboratories>,Albuquerque, NM.

Rulon, J., G. S. Bodvarsson, and P. Montazer. 1986./"P liminary Numerical-Simulations of Groundwater Flow in the Unsaturated/Zone, Yucca Mountain,Nevada. LBL-20553, Lawrence Berkeley Laboratory Bet eley,-..

Runchal, A. K. and B. Sagar. 1989. PORFLO-3: 'A Mathematical Model for FluidFlow, Heat and Mass Transport in Variably Saturated Geoloqic Media, UsersManual - Version 1.0. WHC-EP-0041, Westinghouse Hanford Company, Richland,WA. (to be published)

Sagar, B. and A. K. Runchal. 1989. PORFLO-3: A Mathematical Model for FluidFlow, Heat and Mass Transport in Variabiv.Saturated Geolooic Media. Theorv -Version 1.0. WHC-EP'0042, Westinghouse Hanfor Cmpany, Richiand, WA. (to

be published)

Scott, R. B., and J. Bonk. 1984. Preliminary Geol-o-ic Map of Yucca Mountainwith Geoloaic Sections, Nye County, Nevada. U. S. Geological Survey Open-File Report 84-494, scale 1:12,000.

Scott, R. B., R. W. Spengler, A. R. LappinXand M. P. Chornack. 1983. GeologCharacter of Tuffs in the Unsaturated Zone tYucca Mountain, Southern Nevada

ic

In "Role of the Unsaturated Zone in Radioactive and Hazardous aste Disposal,"edited by J. Mercer, Ann Arbor -Science, Ann Arbor, MI.

Travis, B. J., S. W. Hods6n, HE. Nuttal, T. L. Cook, and R. S. Rundberg.1984. "PreliminaryEstimates of Water Flow and Radionuclide Transport inYucca Mountain," MatC.,Res.N Soc. Symp. Proc., Vol. 26, pp.1039-1047.

U. S. Department of Energy, 1988. Consultation Draft of Site CharacterizationPlan, Yucca-Mountain Site, Nevada Research and Development Area, Nevada.Volume'II, Office of Civilian Radioactive Waste Management, Washington, DC.

U. S. Department of Energy, 1989. Site Characterization Plan, Yucca MountainSuite,]Nevada Research and Development Area, Nevada. Volume VIII, Part B.,Office of Civilian Radioactive Waste Management, Washington, D. C.

van Genuchten, R. 1978. Calculating the Unsaturated Hydraulic Conductivitywith a New Closed-Form Analytical Model. Water Resources Bulletin, PrincetonUniversity Press, Princeton University, Princeton, NJ.

6.2

i .

Wang, J. S. Y., and T. N. Narasimhan. 1985. H drolooic Mechanisms GoverninoFluid Flow in Partially Saturated, Fractured Porous Tuff at Yucca Mountain.SAND84-7202, Sandia National Laboratories, Albuquerque, NM. *

Wang, J. S. Y., and T. N. Narasimhan. 1986. Hydrologic Mechanisms GoverningPartially Saturated Fluid Flow in Fractured Welded Units and Porous 41onweldedUnits at Yucca Mountain. SAND86-7114, Sandia National Laborataries Alquerque,NM.

Wang, J. S. Y., and T. N. Narasimhan. 1988. Hydroloqic Mode ing of Ye icaKand Lateral Movement of Partially Saturated Fluid Flow Near a Fault Zone at.Yucca Mountain. SAND87-7070, Sandia National Laborat 1 - s, Albuquerque, NM>'

-Winograd, I. J., and W. Thordarson. 1975. Hydrogeologic and HydrochemicalFramework, South-Central Great Basin, Nevada-Cal iforwla, vjithfSpecial Referenceto the Nevada Test Site." U. S. Geological Surx'ey Profeslsional Paper 712-C.

.~~~~~V

/~a.I

K

6.3