Assessment of Geology As It Pertains To Modeling Uplift in Jointed Rock: A Basis For Inclusion of Uncertainty in Flow Models

ERDC TR-02-2
Risk Analysis for Dam Safety Program
Assessment of Geology as it Pertains

to Modeling Uplift in Jointed Rock: A Basis
for Inclusion of Uncertainty in Flow Models
William L. Murphy, Robert M. Ebeling, March 2002
and John M. Andersen
Engineer Research and
Development Center
Approved for public release; distribution is unlimited.

The contents of this report are not to be used for advertising,
publication, or promotional purposes. Citation of trade names
does not constitute an official endorsement or approval of the use
of such commercial products.
The findings of this report are not to be construed as an official

Department of the Army position, unless so designated by other
authorized documents.
PRINTED ON RECYCLED PAPER

Risk Analysis for Dam Safety Program ERDC TR-02-2
March 2002
Assessment of Geology as it Pertains

to Modeling Uplift in Jointed Rock: A Basis
for Inclusion of Uncertainty in Flow Models
by William L. Murphy, John M. Andersen
U.S. Army Engineer Research and Development Center
Geotechnical and Structures Laboratory
3909 Halls Ferry Road
Vicksburg, MS 39180-6199
Robert M. Ebeling
Information Technology Laboratory
Final report
Approved for public release; distribution is unlimited
Prepared for U.S. Army Corps of Engineers

Washington, DC 20314-1000
Under Work Unit 6
Contents
Preface ................................................................................................................. xii
Conversion Factors, Non-SI to SI Units of Measurement................................... xiii

1—Introduction ......................................................................................................1
1.1 Background ...............................................................................................1

1.2 Purpose......................................................................................................2
2—Principles of Flow in Jointed Media .................................................................3
2.1 Groundwater Flow and Uplift Pressure in Discontinuous Rock................3

2.1.1 Analytical methods in uplift prediction............................................3
2.1.2 Joint aperture and uplift pressure distribution..................................4
2.2 Coupling Phenomena ..............................................................................12
2.3 Considerations in Assessing Hydraulic Conductivity of Joints ...............13
2.3.1 Permeability and hydraulic conductivity........................................13
2.3.2 Flow in joints: The cubic law ........................................................14
2.3.3 Mechanical and conducting aperture and JRC...............................15
2.3.4 Laminar and turbulent flow in joints .............................................17
2.4 Determination of Rock Mass Hydraulic Conductivity.............................19
2.4.1 Methods using pressure and pumping tests....................................19
2.4.2 Other methods ...............................................................................21
2.5 Approaches in Computing Hydraulic Conductivity of Jointed Rock
Masses...............................................................................................21
2.6 Uncertainty in Determining Hydraulic Properties of Jointed Rock .........23
3—Geological Considerations in Uplift Analysis and Foundation

Investigations .................................................................................................24
3.1 Overview of Geologic Factors.................................................................24
3.2 Rock Type ...............................................................................................26
3.2.1 Characteristics of sedimentary rocks..............................................26
3.2.2 Characteristics of igneous rocks ....................................................27
3.2.3 Characteristics of metamorphic rocks ............................................29
3.2.4 Problems with sedimentary rocks in dam foundations...................29
3.2.5 Problems with igneous rocks in dam foundations..........................30
3.2.6 Problems with metamorphic rocks in dam foundations .................33
3.3 Problems Associated with Weathering....................................................33
3.4 Problems Associated with Stratigraphy...................................................36
iii
3.5 The Role of Discontinuities.....................................................................37
3.5.1 Describing and measuring discontinuities in rock .........................38
3.5.2 Effects of joint condition on groundwater flow .............................42
3.5.3 The problem with faults.................................................................46
3.6 Foundation Investigation Methods ..........................................................47
3.6.1 Surface mapping ............................................................................47
3.6.2 Geophysical explorations...............................................................48
3.6.3 Boring and sampling......................................................................49
3.6.4 Borehole examination and testing..................................................50
3.6.5 Exploratory excavations (tunnels, shafts, drifts, test pits,
trenches) .....................................................................................51
3.6.6 In situ testing .................................................................................51
3.6.7 Laboratory testing ..........................................................................52
3.6.8 Groundwater and foundation seepage investigations.....................52
4—Selection of Case History for Uplift Investigation ..........................................56
4.1 Libby Dam ..............................................................................................57

4.1.1 Description ....................................................................................57
4.1.2 Foundation geology .......................................................................57
4.1.3 Instrumentation ..............................................................................59
4.1.4 Assessment of data quality.............................................................62
4.2 Dworshak Dam........................................................................................64
4.2.1 Description ....................................................................................64
4.3 Green Peter Dam .....................................................................................73
4.3.1 Description ....................................................................................73
4.4 Detroit Dam.............................................................................................80
4.4.1 Description ....................................................................................80
4.5 Wolf Creek Dam .....................................................................................86
4.5.1 Description ....................................................................................86
4.6 Old Hickory Dam ....................................................................................92
4.6.1 Description ....................................................................................92
4.7 J. Percy Priest Dam .................................................................................96
4.7.1 Description ....................................................................................96
4.7.3 Instrumentation ............................................................................101
iv
4.7.4 Assessment of data quality...........................................................101
4.8 Selection of a Case History ...................................................................101
4.9 Other Observations................................................................................105
5—Hydraulic Properties of Jointed Rock from Pressure Tests at Libby Dam ....107
5.1 Background ...........................................................................................107

5.2 Geological Conditions at Libby Dam ....................................................108
5.3 Computation of Equivalent Hydraulic Conductivity .............................111
5.3.1 Principles and assumptions..........................................................111
5.3.2 Libby Dam model ........................................................................112
5.3.3 Pressure tests versus pumping tests .............................................119
5.3.4 Pressure testing at Libby Dam .....................................................119
5.3.5 Results for Libby Dam model......................................................121
5.4 Computation of Parallel Plate Aperture and Joint Hydraulic
Conductivity....................................................................................122
5.5 Discussion of Results ............................................................................137
5.5.1 Equivalent hydraulic conductivity, Ke .........................................137
5.5.2 Conducting aperture, e.................................................................142
6—Analysis of Uplift Pressures at Libby Dam...................................................147
6.1 Calculation of Uplift Pressures from Gauge Readings ..........................149

6.2 Variation of Uplift Pressure with Time .................................................150
6.3 Variation in Uplift Pressure Within the Foundation..............................162
6.4 Significance of Uplift Variations...........................................................165
6.5 Summary ...............................................................................................166
7—Conclusions and Recommendations .............................................................169

7.1 Conclusions...........................................................................................169
7.1.1 Geologic uncertainty....................................................................169
7.1.2 Material uncertainty.....................................................................170
7.1.3 Spatial uncertainty .......................................................................171
7.2 Recommendations .................................................................................172
7.2.1 Geologic data for uplift modeling................................................172
7.2.2 Additional research exercise for existing dams............................175
References ..........................................................................................................176
Appendix A: Explanation of Snow's Equation for Flow Through Fractures ......A1

Appendix B: Derivation of Darcy's Equation of Radial Flow to a Borehole
(Equation 2.11) ............................................................................................. B1
Appendix C: Glossary......................................................................................... C1
Appendix D: Notation ........................................................................................D1
SF 298
v
List of Figures
Figure 2.1. Distribution of uplift pressures in dam foundation for

uniform hydraulic conductivity in a porous medium. ...................5
Figure 2.2. Natural rock joint with mechanical aperture E and

equivalent parallel plates with conducting aperture e . .................6
Figure 2.3. Distribution of uplift pressure in (a) joint of uniform

aperture, (b) two-joint network, and (c) tapered joint ...................8
Figure 2.4. Nonlinear response of uplift pressure (as head) to rising

reservoir headwater at six positions along a continuous
joint of uniform aperture...............................................................9
Figure 2.5. Variation in uplift pressure (head) along (a) uniform,
(b) tapered heel to toe, and (c) tapered toe to heel. e is
joint aperture...............................................................................10
Figure 2.6. Variation in uplift pressure with changes in joint

hydraulic conductivity (changes in joint aperture) with
rising reservoir levels at a point in a dam foundation .................11
Figure 2.7. Hyperbolic model for joint deformation .....................................12
Figure 2.8. Empirical curves relating mechanical aperture (E),

conducting aperture (e), and joint roughness coefficient ............16
Figure 2.9. Range of validity of Darcy's law.................................................19
Figure 3.1. Distribution of major categories of rock in the

conterminous United States ........................................................25
Figure 3.2. Volcanic and metamorphic rocks in the United States................28

Figure 3.3. Geologic map of wall of excavation for Block 21,
foundation of Green Peter Dam ..................................................32
Figure 3.4. Relations between geologic structure in abutment of

Malpasset Dam and arch.............................................................34
Figure 3.5. Geologic section along axis of Red Rock Dam, looking
upstream .....................................................................................37
Figure 3.6. (a) Joint orientation or attitude defined by strike and dip
of joint plane (shaded). (b) Joint system consisting of
joint sets 1 and 2 .........................................................................40
Figure 3.7. Results of finite element joint orientation study .........................41
vi
Figure 3.8. Relationship of fluid conductivity to the distribution of
open fractures in a rock mass......................................................43
Figure 3.9. Influence of joint length on uplift pressure distribution..............44

Figure 3.10. Influence of relative joint aperture and joint inter-
connectivity on uplift pressure distribution.................................45
Figure 4.1. Libby Dam, plan and view looking upstream .............................58
Figure 4.2. Libby Dam, upstream gallery longitudinal section .....................60
Figure 4.3. View upstream of part of upstream grout/drainage

gallery, Libby Dam. ....................................................................61
Figure 4.4. Total head recorded on Gauge P23R1 versus forebay
reading, 1989, Libby Dam ..........................................................62
Figure 4.5. Readings for 6 years from Gauge P23C1, Libby Dam................63
Figure 4.6. Plan view of Dworshak Dam ......................................................65
Figure 4.7. Elevation of Dworshak Dam, looking upstream .........................66
Figure 4.8. Section through Dworshak Dam at station 25+40 ......................67

Figure 4.9. Plan of exploration borings for Dworshak Dam (March
1971) ..........................................................................................68
Figure 4.10. Locations of uplift gauges, Dworshak Dam................................70

Figure 4.11. Detail of instrumentation, monoliths 23 and 25,
Dworshak Dam ...........................................................................71
Figure 4.12. Forebay level versus Gauge P23X reading, Dworshak
Dam ............................................................................................73
Figure 4.13. Forebay level versus Gauge P230 reading, Dworshak
Dam ............................................................................................74
Figure 4.14. Forebay level versus Gauge P230 reading, 1987,
Dworshak Dam ...........................................................................75
Figure 4.15. Plan of Green Peter Dam ............................................................76
Figure 4.16. Sections through Green Peter Dam.............................................77
Figure 4.17. Plan of Green Peter Dam showing locations of borings

and drifts.....................................................................................78
Figure 4.18. Goelogic section along base line, Green Peter Dam ...................79
Figure 4.19. Locations of pressure measurement instruments, Green

Peter Dam ...................................................................................81
Figure 4.20. Plan view of Detroit Dam ...........................................................82
vii
Figure 4.21. Elevation view of Detroit Dam ...................................................83
Figure 4.22. Sections through Detroit Dam ....................................................84

Figure 4.23. Locations of piezometers, Detroit Dam ......................................85
Figure 4.24. Details of instrumentation at Detroit Dam ..................................87
Figure 4.25. Plan view of Wolf Creek Dam....................................................88
Figure 4.26. Sections through Wolf Creek Dam .............................................89

Figure 4.27. Boring locations, Wolf Creek Dam.............................................90
Figure 4.28. Uplift cells in Wolf Creek Dam ..................................................91
Figure 4.29. Plan and section of Old Hickory Dam ........................................93

Figure 4.30. Boring locations, Old Hickory Dam ...........................................94
Figure 4.31. Uplift instruments in Old Hickory Dam .....................................95
Figure 4.32. Plan and section of J. Percy Priest Dam......................................97

Figure 4.33. Sections through concrete portion of J. Percy Priest Dam ..........98
Figure 4.34. Locations of borings, J. Percy Priest Dam ..................................99

Figure 4.35. Geologic section, foundation of J. Percy Priest Dam................100
Figure 3.36. Geologic section along axis of J. Percy Priest Dam..................102

Figure 4.37. Locations of piezometers and uplift gauges, J. Percy
Priest Dam ................................................................................103
Figure 4.38. Plan view of J. Percy Priest Dam showing locations of

uplift gauges .............................................................................104
Figure 5.1. Location of Libby Dam, Kootenai River, Montana ..................109
Figure 5.2. Mapped discontinuities on foundation floor, monoliths

22, 23, and 24, Libby Dam .......................................................110
Figure 5.3. Schematic of borehole pressure tests, Libby Dam ....................114

Figure 5.4. Locations of boreholes, monolith 23 area, Libby Dam .............120
Figure 5.5. Summary log of boring D-92, monolith 23 area, Libby Dam ...121
Figure 5.8. Summary log of boring D-125, monolith 23 area, Libby
Dam ..........................................................................................124

Dam ..........................................................................................125
viii
Figure 5.10. Summary log of boring D-184A, monolith 23 area, Libby
Dam. .........................................................................................126

Dam ..........................................................................................127

Dam ..........................................................................................128
Figure 5.13. Equivalent hydraulic conductivity, number of joints per
test interval, equivalent joint aperture, and equivalent
joint hydraulic conductivity, borehole D-92, monolith 23,
Libby Dam................................................................................129
Libby Dam................................................................................130

Libby Dam................................................................................131
joint hydraulic conductivity, borehole D-125,
monolith 23, Libby Dam...........................................................132


joint hydraulic conductivity, borehole D-184A,

Figure 5.21. Variation in equivalent hydraulic conductivity with

elevation in nine boreholes near monolith 23. From
foundation pressure tests, Libby Dam.......................................138
ix
Figure 5.22. Geologic cross section through foundation of
Figure 5.23. Relationship of equivalent hydraulic conductivity (Ke) of

pressure-tested zones to difference (range) in Ke for test
pairs, Libby Dam ......................................................................141
Figure 5.24. Joint conducting aperture, e, versus elevation, Libby

Dam monolith 23 ......................................................................143
Figure 5.25. Joint conducting aperture, e, versus depth below top of

rock, Libby Dam monolith 23...................................................144
Figure 5.26. Average rock fracture openings (in microns) versus depth
below overburden (top of rock) at dam sites.............................145
Figure 6.1. Instrument and drain locations, monolith 23, Libby Dam ........148
Figure 6.2. Upstream grout and drainage gallery, monolith 23, Libby
Dam, looking upstream.............................................................149
Figure 6.3. Downstream drainage gallery, monolith 23, Libby Dam,

looking upstream ......................................................................150
Figure 6.4. Detail of uplift cell and installation, Libby Dam ......................151
Figure 6.5. Uplift pressure at monolith 23, Libby Dam, L1-5, 1981-
1990..........................................................................................152
Figure 6.6. Uplift pressure at monolith 23, Libby Dam, L1-5, 1991-
1999..........................................................................................153
Figure 6.7. Uplift pressure at monolith 23, Libby Dam, C1-4, 1981-
1990..........................................................................................154
1999..........................................................................................155
1990..........................................................................................156
1999..........................................................................................157
Figure 6.11. Uplift pressure at monolith 23, Libby Dam, R1-5, 1981-
1990..........................................................................................158
Figure 6.12. Uplift pressure at monolith 23, Libby Dam, R1-5, 1991-
1999..........................................................................................159
Figure 6.13. Uplift pressure at monolith 23, Libby Dam, for calendar
year 1990 ..................................................................................161
x
Figure 6.14. Plot of dependence of L1, C1, and R1 (in ft) on forebay
elevation (ft), monolith 23, Libby Dam, for calendar year
1999..........................................................................................162
Figure 6.15. Recorded uplift pressures in 1990 versus design pressures,

Figure 6.16. Recorded uplift pressures in 1999 versus design pressures,

Figure 6.17. Section through drains, monolith 23, Libby Dam.....................167
Figure 6.18. Effect of drains on piezometric surface (idealized)...................168
Figure 7.1. Recommended four-step approach to modeling of uplift

pressures in rock foundations ...................................................172
Figure A.1. Snow’s solid of dimensions W, broken by parallel plane
fractures .....................................................................................A1
List of Tables
Table 2.1 Classifications of Mechanical Aperture ........................................7

Table 3.1 A Classification of the Common Igneous Rocks ........................27
Table 3.2 Classification of Degree of Weathering of Rocks.......................36

Table 3.3 Material Filling Discontinuities and Associated Problems .........42
Table 3.4 Popular Diamond Core Bit Core and Hole Diameters ................49
Table 3.5 In Situ Tests for Rock and Soil...................................................52
Table 3.6 Laboratory Classification and Index Tests for Rock...................53
Table 3.7 Instruments for Measuring Piezometric Pressure........................54
Table 4.1 Summary of Uplift Data Assessment........................................105
Table 5.1 Estimates of Borehole Interval and Joint Hydraulic

Properties from Borehole Pressure Tests, Monolith 23,
Libby Dam................................................................................115
Table 5.2 Statistical Information on Computed Values of

Equivalent Aperture, e..............................................................142
Table 6.1 Pressure Observations in 1990 and 1999..................................165
xi
Preface
This report describes an analysis of geologic factors affecting uncertainty in

the prediction and modeling of uplift beneath concrete gravity dams. This study
was conducted by the Geotechnical and Structures Laboratory (GSL) and the
Information Technology Laboratory (ITL), Vicksburg, MS, U.S. Army Engineer
Research and Development Center (ERDC), in support of Work Unit 6 (Uplift
Uncertainty and Probabilistic Models) of the Risk Analysis for Dam Safety
Program, Mr. H. Wayne Jones, ITL, Program Manager.
Dr. Robert M. Ebeling, ITL, was principal investigator for the work unit. The
report was prepared and written by Messrs. William L. Murphy and John M.
Andersen, GSL, and Dr. Ebeling. Mr. Timothy D. Ables was Acting Director,
ITL, and Dr. Michael J. O’Connor was Director, GSL, during the conduct of this
study.
The authors acknowledge the assistance of the following Corps District and
Division personnel in providing data and guidance for this study: Lawrence
Mann, Steve Meyerholtz, Richard Garrison, Rick Eckerlin, and Lola Schiefelbein,
U.S. Army Engineer District, Seattle; Michelle LeFlore, U.S. Army Engineer
District, Walla Walla; James Griffiths, U.S. Army Engineer District, Portland;
James Gunnels and Jody (John) Stanton, U.S. Army Engineer District, Nashville;
and Dale Munger, U.S. Army Engineer Division, Northwestern. Thanks also to
Vickie Parrish and Gloria Naylor, Computer-Aided Engineering Division, ITL,
for their assistance with report illustrations.
At the time of publication of this report, Dr. James R. Houston was Director
of ERDC, and COL John W. Morris III, EN, was Commander and Executive
Director.
The contents of this report are not to be used for advertising, publication,
or promotional purposes. Citation of trade names does not constitute an
official endorsement or approval of the use of such commercial products.
xii
Conversion Factors, Non-SI to
SI Units of Measurement
Non-SI units of measurement used in this report can be converted to SI units

as follows:
Multiply By To Obtain
cubic feet 0.02832 cubic meters
cubic feet per minute 28.312 liters per minute
degrees 0.01745 radians

feet 0.3048 meters
inches 2.54 centimeters
inches 25.4 mm
inches 25,400 microns
mile 1.609 kilometer

pounds (force) per cubic foot 157.08767 newtons per cubic meter
pounds (force) per square foot 47.88026 pascals
pounds (force) per square inch 6894.757 pascals

square feet 0.0929 square meters
square inches 0.00064516 square meters
xiii
1 Introduction
1.1 Background
Uplift is one of the major forces affecting the stability of rock-founded con-
crete dams. Stability problems associated with uplift can occur for many reasons.
For example, problems often arise during the determination of magnitude and
distribution of uplift pressures and corresponding uplift force within foundations
in heterogeneous rock that contains discrete rock discontinuities. Another problem
is that it is often difficult to extrapolate distributions of uplift pressures within the
rock foundation to reservoir levels above the pool of record. A third example of
problems related to uplift is that if the drains are not maintained over time, their
ability to dissipate uplift pressures diminishes. Consequently, the reliability of the
dam deteriorates with time even if the pool elevation is held constant.
Uplift pressures are controlled by the flow regime within the rock foundation.
The flow regime is a function of site-specific geology. Both the geological inter-
pretation and the analytical procedures used to calculate flow within the founda-
tion introduce uncertainties into the calculation of uplift pressures. A risk
assessment of a dam must account for uncertainties in all factors that impact the
computation of uplift pressures. Currently no methodologies and corresponding
analytical procedures are available for assessing all of the uncertainties in com-
puted uplift pressures in a risk assessment of rock-founded concrete gravity dams.
The goal of research in uplift uncertainty and probabilistic modeling is to

develop a methodology, analytical procedures, and software for the assessment of
uplift pressures and forces within rock foundations for use in the assessment of
the reliability of rock-founded concrete gravity dams. Uncertainties in the geol-
ogy, the flow regimes, and the flow models are to be included in the statistical and
probabilistic mathematical uplift model.
Uncertainty in modeling of uplift pressure in rock foundations manifests itself

in three areas: geologic uncertainty, material uncertainty, and spatial uncertainty.
Geologic uncertainty arises in describing and mapping the stratigraphy, the geo-
logic structure, and the degree of weathering characterizing a foundation. Material
uncertainty pertains primarily to estimates or measurements of rock mass
hydraulic conductivity, particularly from field-test-derived data. Spatial uncer-
tainty is represented in the ways properties vary throughout the foundation and
beyond and between sampling points.
Chapter 1 Introduction 1
1.2 Purpose
The objectives of the study documented in this report were (1) to identify and
characterize geological factors affecting the prediction and modeling of flow and
the development of uplift pressures in rock foundations beneath concrete dams;
(2) to identify the degree and kinds of uncertainty in uplift prediction resulting
from geological investigations of dam foundations, particularly in the description,
testing, and quantification of rock discontinuities; and (3) to select a case history
for assessing the uncertainties associated with geological and uplift analysis of
the foundation of a large concrete dam. This report lays the groundwork for the
development of a systematic characterization of foundation geology in the context
of development of flow models to predict foundation uplift pressures. Numerical
flow modeling will permit prediction of uplift pressures over time and the extra-
polation of uplift pressures to levels above the pool of record.
2 Chapter 1 Introduction
2 Principles of Flow in
Jointed Media
2.1 Groundwater Flow and Uplift Pressure in

Discontinuous Rock
2.1.1 Analytical methods in uplift prediction
Uplift is a major force affecting the stability of concrete gravity dams founded
on rock. Uplift forces decrease the resistance of the dam to sliding. There are
several problems associated with analysis of uplift and its effects on dam stability.
Problems arise when trying to determine the magnitude and distribution of uplift
pressures and resultant uplift forces in foundations that are heterogeneous and that
have discrete rock discontinuities1 (e.g., joints, faults, and bedding planes).
Another problem is extrapolating foundation uplift pressures to pools above the
pool of record. Stability problems arise if drains are not maintained, because the
ability of drains to dissipate uplift pressures diminishes with time. Similarly,
discontinuities may close and become less permeable with time as rising pool
levels compress the foundation, resulting in higher uplift pressures. Nonuniform
stresses imposed by the dam may differentially deform discontinuities, resulting in
tapered joints with varying apertures and variable distribution of uplift pressures.
Conventional equilibrium methods of analysis of the stability of gravity dams

involve assumptions regarding the loading and resisting forces that act on a dam.
Analyses of loading and resisting forces consider the magnitude and distribution
of uplift pressures and effective compressive stresses, respectively, acting normal
to the base of the dam (Pace and Ebeling 1998). Uplift pressure is pore pressure,
defined at any point in the foundation as the unit weight of water (γw) times the
depth below the piezometric surface. Pore pressure reduces the effective stress on
potential failure planes, such as the base of a gravity dam situated on rock, and
thereby lowers the resistance to shear failure along a plane. The relationship is
shown mathematically in the familiar expression:2
1
Throughout this report, for brevity, the term joint or jointed is substituted for the more general
term discontinuity or discontinuous. In many cases, joints are the proper and intended reference. It
should be understood, however, that other discontinuities, including bedding planes, shears, etc.,
are equally pertinent.
2
For convenience, symbols and unusual abbreviations are listed and defined in the Notation
(Appendix D). Engineering and hydraulic terms are defined in the Glossary (Appendix C).
Chapter 2 Principles of Flow in Jointed Media 3

τ = c + (σ - u) tan ϕ (2.1)
where
τ = the shear strength, or shear stress required to cause sliding along a plane
c = cohesion of the rock/rock or concrete/rock interface
σ = the normal stress component of load on the plane
u = the pore (uplift) pressure produced by the head of groundwater
ϕ = the angle of internal friction along the potential failure plane
The term (σ - u) is the effective stress on the plane resulting from the reduction in
normal stress by the pore, or uplift, pressure. In rock foundations, uplift pressures
commonly develop in discrete discontinuities within the rock mass.
Following Stone and Webster Engineering Corporation (1992), uplift pres-

sures can be reported as equivalent piezometric head in feet of water (gauge
pressure in psi, times 2.31 ft1 of water per psi, plus the elevation of the gauge).
Reporting uplift pressures as piezometric head allows comparison directly with
reservoir (headwater) and tailwater elevations. Uplift pressures are controlled by
the flow regime within the rock foundation. The flow regime is a function of site-
specific geology, especially the distribution and geometry of the discontinuities
through which groundwater flows. Both the geological interpretation of founda-
tion conditions and the analytical procedures used to calculate flow within the
foundation introduce uncertainty into the prediction of uplift pressures.
Grenoble et al. (1995) studied the influence of deformation of discontinuities

on uplift pressures in concrete gravity dams (also in Stone and Webster Engi-
neering Corporation (1992), for the Electric Power Research Institute (EPRI)).
They state that stability calculations often assume that the rock mass behaves like
a porous medium and that foundation uplift pressure is distributed linearly from
the upstream face of the dam (or from the position of the drains) to the toe (Fig-
ure 2.1)2. In considerations of flow and developed uplift pressures, the foundation
rock mass cannot be treated as a porous medium unless the joint spacing is so
small that the rock is effectively a continuum. In jointed rock masses, the distri-
bution of uplift pressure is controlled by the geometry and hydraulic conductivity
of the intersecting joints that make up the flow paths beneath the dam.
2.1.2 Joint aperture and uplift pressure distribution
Flow through a joint (or a pipe) is a function of the aperture (size of the
opening) and joint roughness. Joint aperture, discussed further in Chapter 3 of this
report, can be measured in the field with techniques such as borehole camera
surveys. Aperture determines the effective porosity and hydraulic conductivity of
1
A table of factors for converting non-SI units of measurement to metric (SI) units is found on
page xiii.
2
Non-site-specific uplift pressure distribution used in the design and analysis of Corps dams is
given in EM 1110-2-2200 and discussed in Ebeling et al. (2000).
4 Chapter 2 Principles of Flow in Jointed Media

Headwater
H DAM
Tailwater
w H
Commonly assumed linear
uplift distribution
Figure 2.1. Distribution of uplift pressures in dam foundation for uniform hydraulic conductivity in a porous
medium (after Ebeling, Pace, and Morrison 1997)
a jointed rock mass and ultimately affects the distribution of uplift pressure
beneath a dam. Field measurements of rock joints provide what is known as a
mechanical aperture (Barton, Bandis, and Bakhtar 1985). A mechanical aperture
has a degree of asperity, or roughness, manifested by irregularities or undulations
on its surface. Joint roughness affects the flow of water through the joint. Mathe-
matical simulation of flow through the joint requires that the mechanical joint
aperture be reduced to a pair of smooth parallel plates, or a conducting aperture,
for computations of laminar flow and hydraulic conductivity. The mechanical
aperture is designated E and the equivalent, or conducting, aperture is designated

e. The conducting aperture is the distance between two smooth, parallel plates that
would allow the same flow as a mechanical (joint) aperture with rough walls.
Conducting aperture e is always smaller than mechanical aperture E except in the
case of smooth-walled joints. Figure 2.2 illustrates the concept. Mechanical
aperture is measured directly from a rock sample or, in the field, on exposed
joints. Conducting aperture is estimated from permeability or pressure tests.
Rough, natural joint with mechanical aperture E
Smooth, parallel plates with conducting aperture e < E
Figure 2.2. Natural rock joint with mechanical aperture E and equivalent parallel
plates with conducting aperture e (not to scale).
The terms open and tight joints or discontinuities will be used often in this
report. Snow (1968) defined open fractures as those having apertures of 35 µm
(0.35 mm) or greater. His apertures were apparently equivalent, smooth-walled,
conducting apertures computed from borehole pressure tests. Bieniawski (1979),
for his Rock Mass Rating System for tunnel design, considered joints open at
mechanical apertures of 2,500 µm (2.5 mm) or greater. International Society for
Rock Mechanics (ISRM) (1978), proposing discontinuity descriptors for rock
mass classification, defined open joints as those with mechanical apertures of
500 µm (0.5 mm) or greater. Ebeling, Pace, and Morrison (1997) adopted a
mechanical aperture of 250 µm (0.25 mm) as the lower limit of open joints from
work reported in Lee and Farmer (1993), who used data from Barton (1973).
Headquarters, U.S. Army Corps of Engineers (1994), applied a similar classifi-
cation scheme to discontinuities in his discussion of rock mass characterization for
rock foundations. Table 2.1 shows the aperture classification used by Lee and
Farmer (1993) (and Ebeling, Pace, and Morrison 1997) and by Nicholson. The
relative importance of mechanical and conducting apertures to flow modeling and
prediction is discussed further in this chapter and in Chapter 5. Following work by

Table 2.1
Classifications of Mechanical Aperture of Ebeling, Pace, and
Morrison (1997) after Lee and Farmer (1993) and Barton (1973)
µm)1
Mechanical Aperture, mm (µ Class
<0.1 (<100) Very tight
0.10-0.25 (100-250) Tight
0.25-0.50 (250-500) Partly open
0.50-2.50 (500-2,500) Open

2.50-10.0 (2,500-10,000) Moderately wide
>10 (>10,000) Wide

1
Mechanical aperture classification of Nicholson (after Headquarters, U.S. Army Corps of
Engineers, 1994):
(1) Very tight: separations of less than 0.1 mm (<100 µm).
(2) Tight: separations between 0.1 and 0.5 mm (100 µm and 500 µm).
(3) Moderately open: separations between 0.5 and 2.5 mm (500 µm and 2,500 µm).
(4) Open: separations between 2.5 and 10 mm (2,500 µm and 10,000 µm).
(5) Very wide: separations between 10 and 25 mm 10,000 µm and 25,000 µm.
Ebeling and others, a tight joint in this report is assumed to be one with a mechan-
ical aperture less than about 250 µm (0.25 mm).
Grenoble et al. (1995) simulated foundation loading using finite element

analysis and measured uplift pressures on 17 dams over a period of a year. Their
studies showed that rising reservoir levels differentially deformed discontinuities
in the foundation and caused the hydraulic conductivity in rock joints to increase
at the heel and to decrease at the toe, a condition simulated by a tapered joint. If
the hydraulic conductivity does not change, the uplift pressure beneath the dam is
linearly proportional to the headwater pressure (reservoir level). However, if joint
hydraulic conductivity changes because of induced joint deformations (i.e., the
taper of the joint changes with rising reservoir levels), the relationship between
uplift pressure and headwater pressure is nonlinear. The next several paragraphs
discuss the relationship of uplift pressure with reservoir loading. Note that one
relationship is the variation in uplift pressure with distance along the dam base,
and the other is the variation in uplift pressure at a point within the foundation as
headwater pressure changes.
Change in aperture in the direction of flow causes uplift pressure to follow a

curved rather than a linear distribution. Figure 2.3a shows the linear pressure
distribution within a joint of constant aperture from the heel to the toe of a dam.
Figure 2.3b shows the pressure distribution in a joint network represented by a
large-aperture pipe and a small-aperture pipe. Most of the pressure loss occurs in
the smaller pipe because of high frictional losses. The result is, in effect, a non-
linear pressure distribution between the heel and toe. Figure 2.3c shows the
analogy extended to nonlinear pressure distribution for a tapered joint, for which
the aperture changes (steps down) continuously from the heel to the toe.
Changes in loading of the dam foundation, for example by rising reservoir

levels, can decrease joint aperture near the toe and increase joint aperture near the
heel. Pressure against the upstream face of the dam tilts the dam. Deformation of

Figure 2.3. Distribution of uplift pressure in (a) joint of uniform aperture, (b) two-
joint network, and (c) tapered joint (after Grenoble et al. 1995)

joints affects the hydraulic conductivity and the uplift pressures that develop in
the dam foundation. Ebeling and Pace (1996b) and Pace and Ebeling (1998)
investigated the effect of foundation loading (increasing pool elevation) on uplift
pressures using finite element modeling. In their model, they varied the pool
elevation, which resulted in changes in stresses on foundation joints during
reservoir loading (and unloading when the pool was lowered). Figure 2.4 shows
the nonlinear change in uplift pressure (as head) with rising reservoir (headwater)
elevation, measured at six locations along the joint. The nonlinear variation in
uplift head with headwater elevation along the joint reflects the changes in aper-
ture with loading and unloading along the joint. The nonlinear response of uplift
pressure to reservoir height was obtained in what are considered tight joints, i.e.,
joints with mechanical apertures less than about 250 µm (0.25 mm). Stone and
Webster Engineering Corporation (1992) reported that of 17 dams and locks
investigated for uplift stability evaluation, two dams and one lock wall showed a
nonlinear response (in gauge readings) of uplift pressure to reservoir elevation
changes, eight showed linear responses, and six had insufficient data for a
determination.
Figure 2.4. Nonlinear response of uplift pressure (as head) to rising reservoir
headwater at six positions along a continuous joint of uniform aperture
(Ebeling and Pace 1996b)
Ebeling and Pace (1996a) expanded the discussion of the influence of joint
aperture by looking at the effect of tapered joints on the distribution of uplift
pressure across the base of the dam. The direction of the taper influenced distri-
bution of uplift pressure. A joint of uniform aperture across the base of the dam
produced a linear pressure response (Figure 2.5a). A taper with a larger aperture at
the heel than at the toe produced an uplift pressure distribution that is greater than
the conventional linear assumption (i.e., the pressure increased more rapidly along
the length of the dam, Figure 2.5b). A taper with a smaller aperture at the heel
than at the toe of a dam produced an uplift pressure distribution that is less than
the conventional linear assumption (i.e., the pressure increased more slowly along
the length of the dam, Figure 2.5c).

Figure 2.5. Variation in uplift pressure (head) along (a) uniform, (b) tapered heel
to toe, and (c) tapered toe to heel. e is joint aperture (Ebeling and
Pace 1996a or Ebeling, Pace, and Morrison 1997)

Grenoble et al. (1995) illustrated the nonlinearity effect of changing joint taper at
a point in the dam foundation as headwater elevations increased (Figure 2.6). If
joint taper and hydraulic conductivity do not change, the uplift pressure at a point
changes linearly with rising headwaters. If joint taper increases with rising head-
water (joint hydraulic conductivity decreases toward the toe), uplift pressure at the
point increases nonlinearly.
Pressure if joint K decreases

(taper increases) toward the toe
Uplift Pressure
Pressure if joint K (and taper)

do not change
Headwater pressure
Figure 2.6. Variation in uplift pressure with changes in joint hydraulic conductivity
(changes in joint aperture) with rising reservoir levels at a point in a
dam foundation (after Grenoble et al. 1995)
Chapter 3 discusses other geological aspects of discontinuities in controlling

the distribution of uplift pressure in a dam foundation, including joint aperture,
joint length, and joint interconnectivity. Risk assessment of a dam must account
for uncertainties in all factors that impact the computation of uplift pressures. To
compute realistic deterministic or probabilistic estimates of uplift pressures within
a rock foundation, investigative methods must be formulated to permit the devel-
opment of a realistic fluid flow model for a rock foundation with discontinuities.
Stone and Webster Engineering Corporation (1992) provide an excellent evalua-
tion of the effects of geologic conditions on uplift pressure distributions for
several existing large concrete gravity dams.

2.2 Coupling Phenomena
Ebeling, Pace, and Morrison (1997) discussed the relationship between uplift
pressures developed beneath a dam and the stresses imposed on the foundation by
the dam and reservoir. Changes in loading imposed on the rock joints change the
joint apertures, which change the hydraulic conductivity of the joints and ulti-
mately the uplift pressures developed in the foundation. Uplift pressures in turn
affect the stresses imposed by the structure on the foundation. Uplift pressures
developed in the foundation are said to be coupled to the loadings applied by the
structure. Uplift pressures are usually nonlinear since the rock joint aperture varies
across the width of a structure. A tapered joint will produce a nonlinear uplift
pressure distribution, as shown earlier. Coupling phenomena and nonlinearity are
applicable in general to tight joints. A mathematical relationship between the
deformation of joints and the applied loading (or unloading) has been established
from laboratory tests on several different rock types and joints. The deformation
of a joint with applied normal stress is commonly referred to as joint closure or
joint opening. Bandis (1980) modeled joint deformation as a hyperbolic function
applied to jointed rock. Figure 2.7 illustrates Bandis’ hyperbolic relationship
between normal stress (σn) and joint closure (Vj).
Figure 2.7. Hyperbolic model for joint deformation (Ebeling, Pace, and Morrison
1997, after Bandis 1980). σn = effective normal stress, ∆Vj = joint
closure, Vm = maximum joint closure
Ebeling, Pace, and Morrison (1997) observed several relationships from

Bandis’ (1980) joint closure model. The maximum closure (the asymptote, Vm) is
generally 0.3 to 0.9 times the average initial joint aperture. Actual contact areas at
maximum closure generally range from 40 to 70 percent of the total sample area.
The maximum closure for samples with similar initial mechanical aperture
depends primarily on the joint wall compressive strength (JCS) of the rock. The
maximum closure decreases linearly as the joint roughness coefficient (JRC)
increases, irrespective of the JCS. Weathered joints produce larger maximum
closure than unweathered because weathered joints usually have larger initial
mechanical aperture and lower JCS than unweathered joints.

Ebeling, Pace, and Morrison (1997, Chapter 5) used a modified version of
Bandis’ hyperbolic relationship and finite element methods to model the response
of dam foundations to reservoir loading. Their work showed that changes in the
loading of a dam foundation by raising and lowering the reservoir deformed
foundation joints and led to nonlinear changes in developed uplift pressures
beneath the dam.
2.3 Considerations in Assessing Hydraulic

Conductivity of Joints
The preferred method for determining uplift pressures beneath dams is the use
of accurate piezometric instrumentation data. When instrumentation data are not
available or when the reservoir levels to be analyzed exceed those for which the
piezometric measurements were made, other procedures must be used to establish
the distribution of flow and the corresponding uplift pressures. One method
widely used by engineers to establish uplift pressures along a section within a rock
foundation is to compute uplift pressures from flow within rock joints. Ebeling
and Pace (1996a) investigated the fundamentals of flow through jointed rock and
how the dimensions of rock joints, especially joint aperture, influence computed
uplift pressures.
2.3.1 Permeability and hydraulic conductivity
Hydraulic conductivity, K, is the quantifiable ability of a medium to transmit a

fluid under a pressure gradient. Measurements of hydraulic conductivity consider
the properties of the fluid and the medium. Hydraulic conductivity has units of
velocity (length L/time T) but is actually a measure of the volume of a fluid flow-
ing through a cross-sectional area per unit of time under a dimensionless hydraulic
gradient (L3/L2/T). Intrinsic permeability, or simply permeability, k, is a general
term for the ability of a soil or rock to transmit fluid under a hydraulic gradient.
Permeability does not consider the properties of the fluid. Permeability has units
of length squared (L2). Previous literature has variably used the terms coefficient
of permeability and permeability for hydraulic conductivity. Discussions in this
report designate K as hydraulic conductivity and k as permeability, in keeping
with recent usage. K (for water) is related to k by the expression
K = k(γw/µw) (2.2)
where γw and µw are the unit weight and dynamic viscosity of water, respectively.
The relationship of hydraulic conductivity to flow rate (Q) in a porous

medium was established by Darcy as
Q = KiA (2.3)
where

Q = flow rate in volume per unit time
K = hydraulic conductivity of the medium and fluid
i = hydraulic gradient (or δh/δl for one-dimensional flow, where δh/δl is
incremental change in head, h, over length, l)
A = area across which flow occurs
2.3.2 Flow in joints: The cubic law
Using Darcy’s law and substituting an open joint for the porous medium, the
equation for a single joint may be written (Ebeling and Pace 1996a)
Q = Kj⋅i⋅AREAflow (2.4)
where Kj is the hydraulic conductivity of a single joint, and AREAflow (e times unit
width) is the area of flow at any position along the joint. Note that Q is the flow
rate per unit width of the joint.
The joint hydraulic conductivity Kj can be expressed as (Ebeling and Pace

(1996b)
Kj = (γw/12µw)⋅e2 (2.5)
where e is the conducting aperture (in units of length).
Equation 2.5 is derived as follows: intrinsic permeability, k, is expressed in

units of length squared. If d is the pore size of a medium, then k = Cd2, where C is
a dimensionless shape factor (Freeze and Cherry 1979; Davis and DeWiest 1966).
Working from this basic relationship, Snow (1968) showed that for a single planar
joint of aperture e (Appendix A), intrinsic permeability, k = e2/12 (see Appen-
dix A for an explanation of Snow’s equation). Hydraulic conductivity, K, which
considers properties of the fluid and of the medium, is related to k by Equa-
tion 2.2. Substituting Snow’s relationship for k into Equation 2.2 yields Equa-
tion 2.5. Thus, joint hydraulic conductivity, Kj, is proportional to the square of the
conducting aperture, e. Fluid properties (γw/12µw) are introduced into the equation
because Snow’s relationship was derived for intrinsic permeability, k, but it is
hydraulic conductivity, K, that is sought in Equation 2.5.
The cubic law (Ebeling and Pace 1996a) establishes the relationship between
the conducting aperture, e, and Q (as flow rate per unit width) as follows from
Equations 2.4 and 2.5:
Q = (γw/12µw)⋅e2⋅i⋅e (e = area at unit width) (2.6)
or
Q = (γw/12µw)⋅e3⋅i (the cubic law) (2.7)

2.3.3 Mechanical and conducting aperture and JRC
As discussed in this section and in Chapter 5, the roughness of joint walls

controls the ease with which water flows through the joint. Joint roughness also
determines the ratio of mechanical (actual) to conducting (equivalent) aperture.
Rougher or more undulating joint walls result in a higher E/e than do smooth, flat
walls.
A relationship exists between conducting aperture, e, and mechanical

aperture, E (Barton, Bandis, and Bakhtar 1985):
e = (JRC)2.5/(E/e)2 (2.8)
or
E/e = (JRC2.5/e)0.5 (2.9)
where JRC is the joint roughness coefficient of Barton (1973). The equation is for
SI units and is valid only for values of E ≥ e and within a range of aperture of 1 to
1,000 µm (Ebeling, Wahl, and Pace 1997).
Equation 2.8 is an empirical relationship developed by Barton, Bandis, and

Bakhtar (1985) from laboratory and in situ flow tests in real joints and between
smooth cut surfaces. They provide a family of curves for predicting the relation-
ship between JRC, E, and e, using Equation 2.8 (Figure 2.8). The curves imply
several behavioral features. Plane, smooth surfaces with a JRC = 0 have theo-
retical conducting apertures equal to mechanical apertures. Extremely rough joint
surfaces will deviate from E = e (E/e = 1) even at very large apertures. The
rougher the natural joint, the greater will be the ratio of E to e.
Barton described simple tilt tests for determining values of JRC for samples of
jointed rock (Barton 1973; Barton and Choubey 1977). Values of JRC ranged
from 0 for smooth joints to 20 for rough joints with many asperities. Citing
Barton, Bandis, and Bakhtar (1985), Ebeling, Wahl, and Pace (1997) stated that
15 is a typical upper value for JRC.
Barton (1982) suggested another form of Equation 2.8,
e = E2/JRC2.5 (2.10)
with which e could be calculated knowing measured values of JRC and mechani-
cal aperture E. Conversely, E could be estimated knowing the other two variables.
Realistically, however, estimates of conducting aperture, e, are obtained from
steady-state flow tests on isolated joints or zones of joints. The cubic law, Equa-
tion 2.7, provides the relationship for estimating e from flow tests. From constant-
head pressure tests in boreholes, and rearranging Equation 2.7, the equivalent
parallel plate aperture of each joint, e, is calculated using the following
relationships.

Figure 2.8. Empirical curves relating mechanical aperture (E), conducting
aperture (e), and joint roughness coefficient (JRC). Developed from
flow test data (after Barton, Bandis, and Bakhtar 1985)
Darcy’s equation for radial flow to a borehole during a pressure test is1
Q = 2πlKeH/ln (R/ro) (2.11)
where
Q = observed steady-state volume flow rate

Ke = equivalent hydraulic conductivity
H = excess head in test section
R = radius of influence of test
ro = borehole radius
By replacing the length of the test section, l, by the product (Ne) in Equation 2.11,
where N is the number of joints intersecting the test section and e is the equivalent
parallel plate aperture, and replacing Ke by Kj of Equation 2.5 to invoke the cubic
law (Equation 2.7) (Zeigler 1976, Appendix B7), the following expression is
produced:
1
See Appendix B for a derivation of Equation 2.11.

Q = 2πNe(e2γw/12µw)H/ln (R/ro) (2.12)
or
1
 12 µ w  3
 Q ln ( R / ro ) i γ 
e =  w
 (2.13)
 2π NH 
 
Chapter 5 applies these relationships to the estimation of Ke, e, and Kj from

pressure test data for Libby Dam.
2.3.4 Laminar and turbulent flow in joints
Estimates of the hydraulic conductivity of jointed rock masses using Darcy’s

law are valid only for laminar flow. It is important to understand the relationship
between the size of joint openings (aperture), joint roughness, and laminar flow.
Todd (1980) discussed the range of validity of Darcy’s law for flow in porous
media. By analogy with flow in a tube, the Reynolds number (Re) was employed
to define the limit of flows described by Darcy’s law. Experiments showed that
Darcy’s law is valid for porous media, i.e., that flow is laminar, for a Reynolds
number up to about 10.
Ebeling, Wahl, and Pace (1997) discussed the importance of determining

whether flow within a rock joint is laminar or turbulent. The cubic law assumes a
linear relationship between Darcian velocity (or specific discharge) and the
hydraulic gradient and thus is valid only for laminar flow conditions. The
Reynolds number is a dimensionless number expressing the ratio of inertial to
viscous forces in flow. Specifically,
Re = vDh/ν (2.14)
where
Re = Reynolds number
v = mean specific discharge (mean volume rate of flow)
Dh = equivalent hydraulic diameter = 2 times the conducting aperture e (four
times the average flow passage area divided by the perimeter (Iwai
1976))
ν = kinematic viscosity = gµw/γw, where g = acceleration by gravity (Zeigler
1976, p 9)

Zeigler (1976)1 provided the following relationships for laminar, transitional
(nonlinear laminar), and turbulent flow, respectively, in rock fissures:
v = Kj i (laminar flow) (2.15)
vm1 = K′j i (hydraulically smooth regime – turbulent flow

[nonlinear laminar flow]) (2.16)
vm2 = K′j i (rough regime – turbulent flow) (2.17)
where K′j is the turbulent fissure hydraulic conductivity. Equations 2.15 through
2.17 were derived in work by Sharp (1970) and Louis (1969). The exponent m is
generally between 1 and 2 (Zeigler 1976, Appendix B4), with m1 < m2. Equa-
tion 2.16 illustrates that the change in v is not linear with respect to the hydraulic
conductivity in the transitional range.
The higher the Reynolds number, the more likely is turbulence to occur. The
equivalent hydraulic diameter Dh for confined flow in a rock joint is defined as
four times the average flow passage area divided by the perimeter and is equal to
two times the conducting aperture e (Ebeling, Wahl, and Pace 1997, citing Iwai
1976). So Dh = 2e, and
Re = v (2e)/ν, (2.18)
which is the expression for the Reynolds number for flow between smooth
parallel plates.
The critical Re is the Reynolds number at which nonlinear laminar flow starts
to occur. Flow in cylindrical pipes is laminar for Re < 2,100 and turbulent for
Re >> 2,100. For values of Re between 2,100 and 4,000, the flow is transitional
between laminar and turbulent. For open flow in parallel walls, the critical Re is
1,000. For flow in an open channel, the critical Re is 500.
A geometric dependence on Re delineating laminar and turbulent flow also

exists for flow in rock joints. Iwai (1976) showed from studies of flow in rock
joints that turbulent flow was evident when the Reynolds number exceeded a
value of 100. The critical Reynolds number decreases with increasing aperture
roughness. Freeze and Cherry (1979) presented a curve relating flow condition to
specific discharge and Reynolds number (Figure 2.9). The curve shows that
nonlinear laminar flow can occur for Reynolds numbers between 5 and 100.
Nonlinear laminar flow is a transitional state between laminar and turbulent flow.
Laminar and Darcian flow are typical when the Reynolds number is less than
about 5. As Freeze and Cherry state, specific discharge (Darcian velocity, K*i)
and Reynolds numbers are high in wide rock joints.
A key relationship expressed by Ebeling, Wahl, and Pace (1997) is that the
critical Re value decreases with increasing roughness. The point at which linear
laminar flow becomes nonlinear laminar flow is lower in rough joints. Louis
(1969) defined a surface roughness index, S, in terms of roughness and equivalent
hydraulic diameter (equal to 2e):
1
Zeigler (1976) provides a rigorous and thorough review and evaluation of the theory of flow in
fractured rock and of practices for determining rock mass hydraulic properties. He also presents
derivations of most of the equations of flow used in this report.

Figure 2.9. Range of validity of Darcy’s law (after Freeze and Cherry 1979,
Figure 2.28)
S = Rr/Dh (2.19)
where
Rr = height of surface asperities

Dh = equivalent hydraulic diameter = 2e
As Ebeling, Wahl, and Pace (1997) stated, a key aspect of Equation 2.19 is that
joint roughness (as Rr) is related to joint aperture. In addition, the critical
Reynolds number, Re, decreases with increasing surface roughness. The greater
the aperture, the less important to roughness is the height of asperities.
2.4 Determination of Rock Mass Hydraulic

Conductivity
2.4.1 Methods using pressure and pumping tests
Zeigler (1976) summarized field and laboratory methods of determining the

hydraulic conductivity of jointed rock masses. Field tests include injection (pres-
sure) tests, pumping tests, and tracer tests. Laboratory tests include controlled tests
on large specimens representative of the rock mass or on smaller specimens
representing a single discontinuity. Injection or pressure tests inject water into a

borehole or isolated section of borehole under a constant pressure and flow rate.
Injection tests can be conducted with water or with air. Hydraulic conductivity is
related to the flow rate and the hydraulic pressure. Pumping tests commonly
extract water from the rock mass through a borehole (or well). Hydraulic con-
ductivity is computed from observations of well discharge and drawdown in the
well or in nearby observation wells. A tracer test injects a tracer, such as a radio-
isotope, a dye, or a salt solution into a well. Hydraulic conductivity is computed
using the dilution rate or travel time of the tracer to another well or other dis-
charge point. The following paragraphs summarize Zeigler’s review of test
methods.
Water pressure test. In this test, water is pumped into a borehole at constant
flow rate and pressure. Water enters the borehole along its entire length or in an
isolated section sealed off by one or two packers. The test is often called a packer
test or, particularly in Europe, a Lugeon test. Hydraulic conductivity is usually
computed assuming laminar flow into a homogeneous and isotropic medium. The
choice of test equipment and procedures affects the quality of the water pressure
test. An important problem in pressure testing is the loss of pressure caused by
frictional resistance along the flow pipe between the ground surface and the test
section. Despite inherent difficulties, the test has advantages that make it popular.
It is rapid and simple to conduct and, by conducting tests within intervals along
the entire length of the borehole, a conductivity profile can be obtained. The test is
usable above and below the groundwater table. Tests can be conducted in small
boreholes, including the popular NX size. Procedures for conducting water
pressure tests and interpreting the results are presented in Bennett and Anderson
(1982), Zeigler (1976), and Geotechnical Laboratory (1993).
Pressure drop test. The pressure drop test is conducted by pressurizing a

borehole test section to a known value, then stopping the water flow and observ-
ing the rate of pressure drop. The test is usually conducted to supplement a water
pressure test. The pressure drop test requires less water than a pressure test,
making it suitable for areas of limited water supply. The pressure drop test can use
lower initial pressures than the water pressure test and is less likely to cause
widening of discontinuities.
Air pressure test. The air pressure test is similar to the water pressure test but
with air substituted for water as the injection fluid. Flow conductivities computed
from air pressure tests must be converted to water hydraulic conductivity. The air
pressure test has the advantage of a virtually unlimited supply of air for surface
application. However, the conversion of air conductivity to hydraulic conductivity
can lead to erroneous results in certain cases.
Pumping tests. Pumping tests are an established means of determining

hydraulic properties of a large volume of rock mass. Water is pumped from a well,
typically at a constant rate over a certain time period varying from hours to days.
The hydraulic conductivity of the aquifer is related to measured drawdowns in the
well or in observation wells and to well discharge based on assumptions concern-
ing the type of flow (confined, unconfined, or semiconfined), properties of the
aquifer, and flow boundary conditions. Isotropic and anisotropic solutions for
hydraulic conductivity from pumping tests are available. Unlike water pressure

tests, pumping tests are limited to testing strata below the groundwater table.
Pumping tests are difficult to perform in small-diameter boreholes because in-hole
pumps are required. The major disadvantage of pumping tests is the large amount
of time required to conduct the test. Pumping tests evaluate flow in a much larger
volume of the rock mass than do pressure tests.
2.4.2 Other methods
Tracer tests. Tracer tests inject an inert solution (tracer) into an aquifer via a
borehole or well. The dilution rate of the tracer at the injection well (tracer dilu-
tion method) or its travel time to another well (tracer travel time method) can be
used to compute hydraulic conductivity. Test strata may be isolated between
packers to determine a conductivity profile. Radioisotopes, salt solutions, and
fluorescent dyes are commonly used as tracers. Detection of the tracer is by visual
examination of samples or with optical-chemical probes at the detection site.
Tracer tests involve a large portion of the rock mass, thus de-emphasizing the
effects of zones of exceptionally high or low conductivity within the mass. The
tests are rapid and relatively simple to perform, and avoid unnatural conditions
that can result from high injection pressures in other types of tests.
Laboratory tests. Two laboratory tests have been suggested for determining
the hydraulic conductivity of jointed rock (Zeigler 1976). The first measures the
conductivity of a large representative sample, such as those on 1-ft cube blocks
containing more than one discontinuity. The second is to measure the hydraulic
conductivity of a single joint. One investigator studied flow through a single joint
by locking in place upper and lower halves of a rock specimen, measuring the
flow through the joint, and applying the equivalent parallel plate concept to
compute individual joint hydraulic conductivity.
2.5 Approaches in Computing Hydraulic

Conductivity of Jointed Rock Masses
Bennett and Anderson (1982) summarized solutions and approaches
developed by others for computing hydraulic properties of rock masses from
borehole pressure tests. Correct interpretation of pressure test results requires that
assumptions and boundary conditions used in their analysis be valid. Too often,
equations of flow are indiscriminately applied without considering whether the
underlying assumptions and boundary conditions are reasonably satisfied by
actual field conditions.
Two approaches are used to calculate hydraulic conductivity from pressure

tests. The first is the continuum approach. Analysis of flow of an incompressible
fluid (water) through saturated rock or soil is usually made assuming Darcy’s law
to be valid: i.e., there is a linear relationship between hydraulic gradient and flow
velocity. In the continuum approach to analyzing flow in discontinuous, or
jointed, rock masses, flow is assumed to occur uniformly throughout the mass
rather than through individual discontinuities. The hydraulic conductivity

determined in this approach is called the equivalent hydraulic conductivity. The
following conditions must be reasonably met for this approach to be valid:
a. The rock mass is homogeneous, isotropic, and saturated.

b. All flow is radial and symmetric about the borehole axis.
c. The borehole test section is vertical.
d. Flow is steady-state; i.e., equilibrium has been established between the
pumping rate or injection pressure and the head in the rock mass near the
borehole.
e. Flow is laminar; i.e. turbulent flow does not occur.
f. A linear relationship exists between pressure (or gradient) and flow rate;
i.e., Darcy’s law is valid.
g. There is no leakage around the packers that isolate the test section.
h. Inertial terms are negligible; i.e., the change in pressure caused by the
acceleration of flow into the rock mass is negligible.
When conditions a through h are reasonably met, the equivalent hydraulic

conductivity of a pressure-tested zone of rock may be computed from constant-
head (constant-pressure) tests using equations developed by Hvorslev (1951) for
Darcian flow to a well. One form of Hvorslev’s equation was presented as Equa-
tion 2.11 earlier in this chapter. Chapter 5 uses this equation to compute equiva-
lent hydraulic conductivities in the foundation of Libby Dam. Other equations
have been developed for analyses of pressure drop tests and air pressure tests
(Bennett and Anderson 1982 and Zeigler 1976).
In the discontinuum approach to analyzing flow through rock, the mass is

modeled as a system of blocks of low or negligible permeability (hydraulic con-
ductivity) bounded by planar joints having a much higher hydraulic conductivity
than the intact rock mass. The spacing and aperture of all joints intersecting the
borehole test section and the effects of secondary joint systems, those joints that
do not intersect the borehole but do intersect the primary joints, must be
considered. Pressure losses occurring at those intersections and flow occurring
through them can be important in some cases. Flow through fissures (joints) has
generally been modeled using the smooth parallel plate analogy developed by
Snow (1965, 1968). Radial flow governed by Darcy’s law is assumed, and flow is
assumed to occur only through the joints intersecting the borehole test section. For
a discontinuum approach, the parameter for hydraulic conductivity in Darcy’s
equation (Equation 2.11) is replaced by parameters describing the equivalent
parallel plate aperture, e, and the number of joints intersecting the test section
(Equations 2.12 and 2.13), as discussed earlier in this chapter. Finite element
models have also been developed to simulate and analyze flow in jointed media
(for example, Ebeling, Pace, and Morrison 1997).
Bennett and Anderson (1982) state that most references to rock mass analysis
emphasize the homogeneous, isotropic case of hydraulic property distribution and
treat anisotropy as a special condition. They stress, however, that isotropic, homo-
geneous rock is the exception, and anisotropic, nonhomogeneous rock the rule. In

practical design cases, continuum analyses will most likely be used, largely
because of difficulties in defining individual fissure conductivities and structural
orientations of complex fissure networks in the field (Zeigler 1976). Bennett and
Anderson (1982) further suggest that test boreholes are commonly drilled verti-
cally, without regard for rock structure, and pressure tests in them cannot ade-
quately assess directional anisotropy in hydraulic conductivity. Carefully designed
inclined boreholes maximize intersection of other joint sets and permit the contri-
bution of each joint set to overall rock mass hydraulic conductivity and seepage.
Another method of assessing anisotropy is the installation of additional vertical
boreholes at different bearings from the vertical test hole and within the zone of
influence and monitoring pressure differences within the monitoring holes.
2.6 Uncertainty in Determining Hydraulic

Properties of Jointed Rock
Uncertainty arises in measuring or estimating E and e, in measuring asperity
or roughness in the field, and in extrapolating measurements in boreholes or
surface exposures to the entire foundation rock mass. Equations for modeling of
flow in jointed rock require values of conducting aperture. Because joint apertures
change with distance from the measuring point, such as a borehole, measured E
may not accurately describe joint widths within the conducting rock mass. Mea-
surements of mechanical apertures, E, must be converted to conducting apertures,
e, for modeling purposes. There is a great deal of uncertainty in estimating con-
ducting aperture from measured mechanical apertures because of the nonlinear
relationship between real and conducting aperture and the dependence on a
measurement of asperity or JRC. Extrapolation of measured joint features from
the measuring point to the rest of the rock mass further weakens the relationship.
Back-calculation of e from aquifer testing is susceptible to errors in measurement
precision, validity of assumptions made to compute aquifer properties from
Darcy’s law, accuracy of joint count in the borehole log, the introduction of non-
linear or laminar flow during testing, and the possibility of artificially widening
joints by hydrofracturing.
Another aspect of flow in jointed rock that is difficult to monitor or ascertain

during foundation investigation is persistence of jointing throughout the rock
mass. Joints may tighten or cease at some unknown distance from the observation
point. Or they may become wider, and may introduce nonlinear or turbulent flow.
Joints may or may not intersect other joints. It is always difficult to make a three-
dimensional assessment based on a limited number of one- or two-dimensional
measurements.
The effects on the joint system with time must also be considered. Changing
reservoir levels deform the foundation through the moment-induced stresses
resulting from reservoir loading of the upstream face of the concrete dam and
through pore pressures developed within the joint system. There are practical
difficulties associated with making in situ measurements beneath an existing dam.
Analytical procedures, such as the finite element method of analysis, may be used
to gain insight into the magnitude and geometry of these changes.

3 Geological Considerations
in Uplift Analysis and
Foundation Investigations
3.1 Overview of Geologic Factors

Uplift pressure develops in the saturated rock mass beneath gravity dams as a
result of the head imposed by the reservoir. Foundation treatment must both
minimize the leakage of reservoir water beneath the dam and manage the uplift
pressures that develop from restricted flow. As Deere (1981) stated:
“Full reservoir pressure imposes severe performance requirements on

the … foundation rock of a high concrete dam. The rock must resist the
bearing and shear stresses transmitted by the dam as well as the uplift
and seepage stresses associated with seepage flow beneath and around
the dam. [At the same time,] the quantity of leakage must also be of
reasonably small magnitude.”
Flow of water in most rock is in the discontinuities that exist within the rock
mass. Stone and Webster Engineering Corporation (1992) suggest that the ideal
foundation with respect to uplift pressures is one that is impervious immediately
below the heel of the dam and free-draining downstream. Under these ideal condi-
tions, the impervious cutoff blocks most of the seepage, and the free-draining
zone allows any seepage water bypassing the cutoff to flow freely to the tailwater
and prevents the development of excessive uplift pressures.
Some rocks have a primary porosity of interconnected pores between grains

that conducts water under pressure. However, most groundwater flow in rock
occurs through the network of discontinuities cutting the rock mass. The role of
flow and pressure in discontinuities is discussed in detail later in this chapter.
Other features of rock affect the stability and hydraulic characteristics of rock
foundations. The kind, or classification, of rock present in the foundation may
affect the velocity of groundwater flow and the distribution of pressure within the
rock mass. Different rock types present different problems in foundation investi-
gation and in control of uplift pressures following reservoir filling. Geologic
classification commonly recognizes three broad divisions of rock: sedimentary,
igneous, and metamorphic. Igneous rock is further distinguished as intrusive or
extrusive, or volcanic. Figure 3.1 shows the distribution of the major categories of
24 Chapter 3 Geological Considerations in Uplift Analysis and Foundation Investigations

Chapter 3
Geological Considerations in Uplift Analysis and Foundation Investigations
25
Figure 3.1. Distribution of major categories of rock in the conterminous United States (data from Schruben et al. 1997)
rock in the United States and where these rock types appear at or very near the
ground surface.
Weathering, the disintegration and decomposition of rock in place by mechan-

ical and chemical processes, weakens the rock mass and increases rock mass
hydraulic conductivity by widening discontinuities. Rock is attacked by weather-
ing agents on exposed surfaces, such as excavation walls and natural outcrops,
and along joints and other discontinuities that extend into the rock mass. The zone
of weathering is most pronounced near discontinuities and exposed surfaces. The
degree of weathering in a rock mass depends on (a) the area of exposed surface,
(b) the age of the exposed surface, (c) the extent of access to the rock mass along
discontinuities and through pores, (d) the chemical composition (mineral content)
and texture of the rock, (e) the environment or climate, and (f) the position and
chemistry of the ground water.
Site stratigraphy, which describes the sequence of rock layers occurring in the
foundation, can create considerable anisotropy in the foundation when strata alter-
nate with depth. A site with variable stratigraphy is more difficult to investigate
because a greater number of samples must be taken to accommodate the changing
rock properties. There is a degree of uncertainty in determining when enough
samples and a sufficient number and spacing of borings, adits, or other sampling
accesses have been achieved. A monolithic foundation can be evaluated with a
minimum of borings and samples. The foundation picture is complicated even
more when site stratigraphy is disturbed by geologic structure. Sedimentary units
originally deposited horizontally may be folded or tilted so that units occurring at
a depth in one part of the foundation will occur deeper or shallower in another
part because of dip. Faults may displace the entire sequence laterally or vertically
by feet or tens of feet, making correlation of strata across the site difficult. Where
a fault brings nonpervious beds adjacent to pervious strata at the fault contact,
foundation drainage may be blocked, leading to buildup of pore and uplift
pressures.
3.2 Rock Type

3.2.1 Characteristics of sedimentary rocks
Sedimentary rock consists of particles that have been eroded and deposited
originally as soft materials and have hardened and gained strength through time.
Compression of deposits by accumulating thicknesses of sediments further lithi-
fies and indurates the material. Sedimentary rock is commonly subdivided into
chemical, organic, and mechanical deposits. Chemical deposits form by materials
precipitating out of water solution or through evaporation of salt-laden bodies of
water. Examples of chemically derived sedimentary rocks are limestone, gypsum,
and anhydrite. Organic rock forms from deposits of vegetative and animal
remains. Coal is an example of an organic sedimentary rock. Mechanically
deposited rock, also called clastic sedimentary rock, consists of accumulations of
fragments of older rocks ranging from clay to boulder-sized particles. Clastic
sediments are carried and deposited by wind, water, ice, or gravity. Examples of
clastic sedimentary rocks are shale, sandstone, and glacial till.

3.2.2 Characteristics of igneous rocks
Intrusive rocks are derived from molten magma injected into the earth’s crust
where it cools slowly enough to form visible crystals within the rock. Examples of
intrusive rocks are granite, syenite, and diorite. Extrusive rocks are derived from
molten magma ejected into the air or onto the earth’s surface. They form volcanic
deposits of flows (for example, basalt or rhyolite) or pyroclastics (for example,
tuff or agglomerate). Intermediate forms of igneous rocks with textures containing
both fine and coarse crystals also occur. Table 3.1 is a classification of the com-
mon igneous rocks. Igneous intrusive rocks typically occur in mountainous
regions, where they often form the core of mountain ranges. Areas of igneous
intrusives include the Appalachians, the Adirondacks of upstate New York, the
Rockies, the Sierras of California, and the northern Cascades of Washington. Pre-
Cambrian granitic intrusives occupy much of northern Minnesota. Most of the
volcanic rocks in the United States occur in the geologically active western states,
especially Washington, Oregon, Idaho, California, Nevada, and Arizona. The
great outpourings of lava flows in the western states are generally very young, of
late Tertiary age.
Figure 3.2 shows the distribution of volcanic and metamorphic rocks in the
conterminous United States.
Table 3.1
A Classification of the Common Igneous Rocks
Light-colored (acidic) Intermediate in color
Quartz- Dark-colored
Texture Quartz-rich deficient Quartz-rich Quartz-deficient (Basic) Ultra-basic1
Gradational change depending on kind and amount of feldspar present
QUARTZ
Coarse texture MONZONITE
PERIDOTITE,
(plutonic, (GRANODIORITE) MONZONITE to
GRANITE SYENITE GABBRO PYROXENITE,
intrusive, to QUARTZ DIORITE
DUNITE, etc.
phaneritic)2 DIORITE
(TONALITE)
Contrasting GRANITE MONZONITE

texture PORPHYRY PORPHYRY
DIABASE OR
(nonuniform
DOLERITE
grain size, or RHYOLITE LATITE
porphyritic) PORPHYRY PORPHYRY
RHYOLITE3
Fine texture or
(OBSIDIAN
glassy (volcanic, QUARTZ LATITE to LATITE to
is a glassy TRACHYTE3 BASALT4
extrusive, DACITE3 ANDESITE3
form of
aphanitic)
rhyolite)
1
Composed wholly of dark minerals, often only one mineral.
2
A pegmatite is a very coarse-grained rock, with most grains larger than 1 cm; some may exceed 1 m in diameter.
3
Light-colored, fine-grained igneous rocks are sometimes called FELSITE.
4
Dark-colored, fine-grained igneous rocks are sometimes called TRAPROCK.
Chapter 3 Geological Considerations in Uplift Analysis and Foundation Investigations 27

Figure 3.2. Volcanic and metamorphic rocks in the United States (data from Schruben et al. 1997)

3.2.3 Characteristics of metamorphic rocks
Metamorphic rocks are rocks that have been altered by heat and pressure,
usually as a result of mountain-building processes or tectonic plate interaction.
Most pre-Cambrian sedimentary rocks older than 800 million years have been
metamorphosed and are called metasediments. The metasediments have been
sufficiently altered by metamorphic processes of heat and pressure that they have
lost their original sedimentary structure, fabric, or texture. For example, sand-
stones have been altered to hard quartzites. Other examples of metamorphic rocks
are marble, slate, schist, and gneiss. Metamorphic rocks occur in areas of moun-
tain building, for example in the Appalachians and Rockies, and in areas of
tectonic rifting and regional faulting, as in southern California and Arizona. They
also are found in the very old, pre-Cambrian rocks of central and northern
Minnesota where part of the Canadian Shield is exposed (see Figure 3.2).
3.2.4 Problems with sedimentary rocks in dam foundations
Goodman (1990) summarized the properties of different rock types with

respect to dam design and construction. Sandstone and other coarse-grained
clastic rocks may exhibit two kinds of hydraulic conductivity. These rocks are
porous to varying extent and may be pervious to groundwater if the porosity is
high enough and the pores are connected. Intergranular porosity is often called
primary porosity. Clastic rocks are commonly jointed and have bedding planes
along which flow can occur. These discontinuities make up the secondary porosity
in the rock mass. If intergranular porosity is low or if the pores are partially filled
with cementing materials1 or fines, the secondary porosity of the discontinuities
will dominate groundwater flow. Goodman stresses the need for field conductivity
testing to determine the hydraulic conductivity and other engineering character-
istics of clastic rock masses. Some sandstones can be grouted effectively, but very
fine-grained or clayey sandstones may be hard to seal by grouting. Foundation
exploration must be able to determine the stratigraphy and structure (the mor-
phology) of the sandstone bodies. Some sandstone deposits, particularly alluvial
or deltaic formations, are of limited lateral extent and may thin, thicken, or dis-
appear in a short distance. Hard, cemented sandstones tend to have open joints
that can produce excessive seepage beneath the dam.
Shales and other fine-grained clastic sedimentary rocks need special attention
in foundation investigations. Shales (as well as mudstones and claystones) may be
expansive because of particular clays composing them, or they may have a pro-
pensity to slake, to disintegrate with changing moisture conditions. Foundation
investigations should strive to describe these rocks adequately. Shaley or clayey
rocks usually have very low hydraulic conductivity because of their fine-grained
nature and clay content. Some shale formations, however, have well-developed
joint networks in the field and may be quite pervious. Field conductivity tests are
preferred to laboratory tests of individual samples because the latter will overlook
the effects of the joint network on hydraulic conductivity of the rock mass.
1
The grains of sandstones and siltstones are often partially cemented together by chemicals
brought out of solution, such as calcium carbonate (calcite) and silicon dioxide (silica).

Stone and Webster Engineering Corporation (1992) described the effects of a
prominent shear zone in the shale foundation of Morris Sheppard Dam in Texas.
The shear zone was parallel to the near-horizontal bedding and ran through the
foundation above the dam to downstream of the spillway and in fact had caused
the spillway to slide. Investigations indicated that the shear zone was acting as a
conduit, distributing water under high uplift pressure beneath the base of most of
the spillway. The uplift pressure was eventually controlled by drilling 147 drains
into the shear zone near the heel.
Shales and sandstones often occur together in alternating, interbedded, and

cyclical layers. This poses a special problem because flow is restrained to the
sandstone layers, and pressures in each sandstone bed can be different from
pressures in other beds isolated by shale layers. Flow velocities in the sandstone
layers may be higher because the area of flow is reduced by the presence of the
impermeable shales. Higher flow velocities can lead to piping or erosion of
weakly cemented sandstones if not controlled.
Limestones and other precipitate rocks may be soluble in water, particularly

along planes of weakness, such as bedding and joints. Groundwater flow paths
along these discontinuities may widen to the point of producing conduit flow with
high velocities. Solution-formed cavities and caves commonly form. The primary
concern for dams constructed on soluble rock is, of course, the loss of reservoir
retention by excessive seepage through the foundation. But uplift pressures may
vary greatly within the foundation, depending on how the system of joints and
bedding planes interconnects with solution cavities.
Douglas Dam is a 203-ft-high concrete gravity dam founded on solution lime-

stone in Tennessee. Stone and Webster Engineering Corporation (1992) described
conditions in the foundation of the dam. Major solution features (sinks, conduits,
and caves) existed beneath the river at the dam site prior to excavation. An exten-
sive weathered zone of highly permeable limestone extended to a depth of over
200 ft below the channel. To treat these conditions, the solution cavities were
cleaned and filled with concrete. The deeply weathered zone was treated by an
extensive grouting program. Uplift pressure measurements after dam construction
showed piezometric levels beneath the dam to be lower than the tailwater. Precon-
struction remediation of the potentially troublesome karst features was effective in
controlling seepage and in reducing uplift pressures.
3.2.5 Problems with igneous rocks in dam foundations
Intrusive igneous rocks have little or no primary porosity for conveying

groundwater. They are commonly jointed, however, and can have substantial
fracture porosity. Weathering of near-surface igneous bodies breaks up mineral
grain boundaries and increases hydraulic conductivity in the weathered zone.
Intrusive rock masses are commonly bounded by irregular contacts. It may be
difficult during field exploration to determine the limits of intrusive rock bodies.
They may have a regular system of joints reflecting regional tectonic stresses.
Many igneous rock bodies also have a subhorizontal set of open, stress-relief
joints near the ground surface caused by weathering and unloading of the rock

mass by erosion of overlying rock. This phenomenon is also called sheeting. The
effects of stress relief and weathering decrease with depth, so joint frequency and
aperture usually decrease with depth in the foundation.
Areas of volcanic rocks cause special problems in foundation investigations.

The texture of volcanic rock can vary within a construction site from hard, dense
lava flows to soft tuffs (deposits of volcanic ash). Volcanic deposits can be
parallel-bedded similarly to sedimentary rocks, or massive and irregularly shaped,
with lateral changes from one rock type and texture to another over a short
distance. Hard lava rock may overlie older but softer tuff deposits at depth.
Volcanic deposits can be very heterogeneous and difficult to describe adequately
during foundation investigations. Groundwater is conveyed through pores in
pyroclastic materials and some vesicular basalts, along contacts and shear zones
between deposits, and through joints developed in the more brittle rocks.
The foundation of Green Peter Dam, in west-central Oregon, illustrates the

complexity associated with volcanic sequences. Figure 3.3 is a geologic map of
the wall of the excavation for Block 21 of the foundation of Green Peter Dam.
The foundation is in dipping, alternating layers of basalt flow, flow breccia, and
tuff. The rock mass is cut by basalt intrusives (dikes). Shear zones of highly
fractured rock developed in the tuff beds. Faults and joints persist throughout the
foundation rock mass. Water flows from the planar features varied considerably
from feature to feature and from place to place within the foundation. Some of the
shear zones had to be accessed through drifts blasted into the foundation rock.
Foundation remediation methods included rock bolting, presplitting and line
drilling to prevent overbreakage, contact grouting, emplacement of a grout
curtain, drainage tunnels, and a drainage curtain.
Stone and Webster Engineering Corporation (1992) gave an example of

foundation treatment designed to control uplift pressures in layered basalts
beneath McNary Dam in southeastern Washington. The foundation consists of
roughly horizontal basalt flows approximately 100 ft thick separated by an
interbed of lacustrine (lake-deposited) sediments about 50 ft thick. The upper
several feet of basalt are severely jointed and pervious. The lower basalts are
columnar-jointed with moderate hydraulic conductivity. The sedimentary interbed
is tuffaceous, generally fine-grained and clayey, and is relatively impervious.
Corps District personnel constructed a grout curtain from a gallery near the heel of
the dam. The grout holes were bottomed in the impervious sedimentary interbed
to form a cutoff wall. Monitoring of piezometric pressures in drain holes down-
stream of the grout curtain indicated little increase in uplift pressures. The rock
was sufficiently permeable to allow seepage to drain freely to the tailwater.
Measurements in adjacent massive basalts, which lacked open, interconnected
joints, revealed much higher uplift pressures.
A particular kind of joint unique to extrusive rocks such as basalt is columnar

jointing. Columnar joints form near vertically at right angles to the direction of
flow of a lava body, in roughly hexagonal prisms. The columns form as a result of
cooling of the lava and stress relief. Columnar-jointed lavas can be very perme-
able to groundwater. The lava flow containing the columnar joints, however, is
limited in extent. Foundation investigation should define the limits of the flow to

32
Chapter 3
Figure 3.3. Geologic map of wall of excavation for Block 21, foundation of Green Peter Dam (after U.S. Army Engineer District, Portland 1969)
Geological Considerations in Uplift Analysis and Foundation Investigations

allow a determination of where in the foundation high flow zones will occur and
where they may be curtailed by adjacent rocks.
3.2.6 Problems with metamorphic rocks in dam foundations
Metamorphic rocks can be especially problematic for dam foundations in a

number of ways. These rocks often occur in complex assemblages with widely
varying properties. Because of their usually great age and their history of forma-
tion in regions of high stress, they commonly contain joints, faults, foliation, and
shear zones. These are planes of weakness along which failure can occur and
along which groundwater can move. Metamorphic rocks are sometimes severely
folded and contorted, forming complex failure and flow paths. Metamorphism
often forms schistose zones characterized by weak, platy minerals aligned in a
direction parallel to bedding. Schistose zones have a low resistance to movement
and comprise potential failure paths in an excavation or foundation.
Goodman (1990) provided an example of dam failure attributed to uplift

pressures developed in metamorphic rock at Malpasset Dam. A wedge of rock
was created by the intersection of a steeply inclined foliation shear and a fault
under the left abutment. When the reservoir filled, water pressure in the schistose
rock rose because joints in the rock closed and prevented free drainage through
the abutment. The wedge of rock yielded under the thrust of the reservoir and dam
and the elevated uplift pressures in the joints. A diagram presented by Bellier
(1977) illustrates the adverse orientations of foliation and structure that probably
led to the Malpasset failure (Figure 3.4). The diagram is a plan and sections
through the left and right abutments of the arch dam. Failure occurred in the left
abutment (circled on the diagram). Bellier explained that the thrust of the arch
into the left abutment was parallel to the direction of foliation in the metamor-
phosed rock. Thrust stresses were concentrated in a narrow section of the abut-
ment bounded by the arch and a downstream fault. The subsurface “dam” thus
formed prevented drainage of groundwater as the reservoir filled. Thrust of the
arch further worsened the condition by compressing foliation joints within the
isolated zone and reducing their hydraulic conductivity. Investigators estimated
that the hydraulic conductivity of the zone decreased by a factor of 100 to 1,000.
Pore pressures increased on the downstream face of the slice. The resultant force
of the arch thrust and the elevated pore pressures produced an ascending move-
ment of the left abutment rock, as shown by the dashed arrow in Section BB of
Figure 3.4.
3.3 Problems Associated with Weathering

Weathering, the process of breaking down rock in place by mechanical and
chemical means, changes the engineering characteristics of rock. Mechanical, or
physical, weathering of rock occurs by many processes. Water that seeps into
pores and open discontinuities, particularly in temperate climates, may freeze and
expand, breaking the rock and rock grains by frost wedging. Thermal expansion
and contraction from severe daily temperature variations, especially in arid
regions, cracks rock or causes it to spall and disintegrate. Clay-rich rocks can be

Figure 3.4. Relations between geologic structure in abutment of Malpasset Dam and arch. Circles indi-
cate left abutment conditions. Dashed arrow (added) in Section BB is direction of uplift force.
See text for explanation (after Bellier 1977)
broken down by cycles of wetting and drying. Unloading is the release of stresses
within the rock mass by erosion of overlying or surrounding materials or glacial
melting and retreat, or by blasting or other excavation of the surrounding rock
mass. Unloading affects all rocks and causes widening of existing discontinuities
such as bedding planes and joints or opening of incipient fractures. Root growth
can affect all rocks and causes widening of discontinuities and eventual fragmen-
tation of the rock mass. The actions of animals, especially burrowing animals, are
most pronounced in the softer sedimentary rocks.
Chemical weathering occurs by the reaction of water, acids and bases, oxy-
gen, and carbon dioxide with the mineral constituents of the rock. Iron sulfides
combine with oxygen to form the commonly occurring red oxides of iron by the

process of oxidation. Carbon dioxide dissolved in water readily dissolves soluble
carbonates such as limestones and dolomites to produce the networks of caves and
solution-enlarged discontinuities of karst regions. Many clay minerals, which are
significant in the stability of geotechnical structures, form from silicates of the
igneous rocks by the addition of water. This process is called hydrolysis, and
under certain conditions it forms hydrous compounds. For example, feldspars, the
most common igneous minerals, alter in the presence of water to illite or kaolinite,
common clay minerals. Oxidation and reduction alter rock minerals to form
oxides or hydroxides. Dissolution, or water solubility, occurs in certain sedimen-
tary rocks including evaporites (e.g., gypsum) and limestone. Part of the rock is
dissolved and carried away, leaving cavities, a soft residue, or weakened zones
within the rock mass.
The absorption of free water into the mineral structure by hydration also
produces a kind of mechanical weathering by expansion of the structure when a
mineral undergoes growth by recrystallization. For example, the hydration of
anhydrite (the “dewatered” form of calcium sulfate) to reform gypsum (hydrous
calcium sulfate) produces a volume change of as much as 30 to 60 percent
(Robinson 1982). The most significant hydration occurs in the alumino-silicate
minerals of igneous rocks (Schultz and Cleaves 1955). The volume increase that
accompanies hydration is an important factor in the disintegration of coarse-
grained igneous and metamorphic rocks. Clay minerals, such as montmorillonite,
that fill the space between discontinuity walls may absorb water and contribute to
expansion and mechanical breaking of the rock mass.
The weathering process occurs worldwide. It proceeds faster at the exposed

rock surface and along permeable discontinuities, particularly joints and fault
zones. It is important to define the depth of the weathered zone and to be prepared
to address problems associated with weakened rock and enlarged water passages
caused by weathering. The weathering profile from the surface downward is trans-
itional. The degrees of weathering and of deterioration of the rock mass decrease
with depth. The transitional zone has been defined by Deere (1981) as the grada-
tional change from residual soil to weathered rock. The depth of the transitional
zone varies with rock type, climate, and other factors and proceeds preferentially
along planes or zones of discontinuity within the rock mass. In dipping strata,
weathering proceeds preferentially and more deeply along beds more susceptible
to chemical weathering. Rock mass properties, including strength and hydraulic
conductivity, vary extensively within the transition zone. The variation in depth,
the irregularity of the weathered zone, and the accompanying range of rock mass
properties account for much of the uncertainty in investigating, preparing, and
treating excavated foundations for concrete dams.
The weathering of rock is recognized by a decrease in the luster of the rock’s

minerals, discoloration of the rock, separation of rock crystals or grains along their
boundaries, increased friability (crumbling when rubbed between the fingers), and
a general decrease in competency and compressive strength. Infiltrating water may
stain discontinuity walls or introduce material to fill open discontinuities. The
degree of weathering can be classified on the basis of simple qualitative visual and
physical inspection. Table 3.2 is a classification developed by Bieniawski (1979)
to describe discontinuities in rock masses.

Table 3.2
Classification of Degree of Weathering of Rocks (after Bieniawski
1979)
Classification Description
Unweathered No visible signs of weathering; rock fresh; crystals bright.
Slightly weathered Discontinuities are stained or discolored and may contain a thin
filling of altered material. The discoloration may extend into the rock
from the discontinuity surfaces to a distance of up to 20 percent of
the discontinuity spacing.
Moderately weathered Slight discoloration extends from discontinuity planes for greater
than 20 percent of the discontinuity spacing. Discontinuities may
contain filling of altered material. Partial opening of grain boundaries
may be observed.
Highly weathered Discoloration extends throughout the rock and the rock material is
friable. The original texture of the rock generally has been
preserved, but separation of the grains or crystals has occurred.
Completely weathered rock The rock is totally discolored and decomposed and friable. The
external appearance of the rock sample is that of soil. Internally, the
rock structure is partially preserved but grains and crystals have
completely separated.
3.4 Problems Associated with Stratigraphy

The overview for this chapter explained that problems with stratigraphy in
dam foundations arise primarily from the variation in rock properties across the
site and with depth. Sedimentary rocks, in particular, can be stratigraphically
diverse. Changes in depositional environments, from deep sea to terrestrial, occur
over time, with consequential variations in the types of sediment deposited.
Limestones may give way with depth to shales or sandstones, and then the cycle
may repeat. In deltaic and near-shore environments, both lateral and vertical
changes in sediments occur repeatedly with time as sea level rises and falls and
rivers change course as they meander toward the sea. The result is a complex
assemblage of sinuous or isolated sand channels, extensive backswamp mud and
clay deposits, clay- or silt-filled abandoned channels, overbank deposits, and
deltaic fans. These deposits show great variation in shape, size, thickness, and,
when converted over time to rock, physical properties. This variation in properties
laterally and with depth causes uncertainty in how a foundation will respond to
changing stress conditions, how and where it will conduct water, and how and
where uplift pressures will develop and be distributed.
Foundation investigations for Red Rock Dam on the Des Moines River in
Iowa encountered some of these problems. Red Rock Dam is not a concrete
gravity structure, but its foundation stratigraphy exemplifies conditions that could
apply in other dam situations. The dam is founded in Mississippian-aged rock of
alternating strata of shale, sandstone, and limestone. Figure 3.5 is a geologic
section along the dam axis, looking upstream. Alternating sandstone and lime-
stone units, generally 10 to 15 ft thick, persist across the foundation to a depth of
about 70 ft below the channel bottom. In some areas, domes of gypsum occur at
the base of this sequence (see Figure 3.5). Updoming of the gypsum beds
intensely fractured the overlying sandstone and limestone strata. Fracturing
allowed river water to seep into the gypsum beds and remove part of the gypsum

800 TOP OF DAM
750
SHALE
ELEVATION, FT
SHALE
ORIGINAL GROUND SURFACE

700
SHALY LIMESTONE
SANDSTONE ORIGINAL TOP OF ROCK
LIMESTONE
650 SANDSTONE
LIMESTONE
SHALY LIMESTONE
GYPSUM
DOLOMITIC HARD, CHERTY SHALE

620
0 250 500
HORIZONTAL SCALE, FT
Figure 3.5. Geologic section along axis of Red Rock Dam, looking upstream (after U.S. Army Engineer
District, Rock Island, 1965)
by dissolution, resulting in subsidence of the overlying rock into the cavities.

Infiltrating water leached the relatively friable sandstone and caused intense
weathering along joints. The limestone was affected by dissolution along joints
and bedding planes and filling of the dissolution cavities by residue and clay.
Hydraulic gradients varied considerably from one side of the foundation to the
other. Concerns with excessive seepage and continued dissolution and piping of
the limestones and gypsum beds, which could compromise the stability of the
dam, required extensive remedial grouting.
3.5 The Role of Discontinuities

Dam foundation investigations for uplift prediction should assess those
geologic conditions controlling groundwater flow beneath the dam. Among the

most important geologic factors are the persistence and condition of rock mass
discontinuities. Murphy (1985) reviewed descriptive terminology and classifi-
cation schemes for rock masses, assessed the importance of rock mass descriptors
in geotechnical applications, and recommended descriptors for critical rock mass
properties and characteristics. The term discontinuity encompasses all natural
breaks or planes of separation in a rock mass. Discontinuities include structural
features, such as faults, joints, and shears; depositional features, such as bedding
planes; post-depositional features, such as schistosity and foliation; and solution-
widened passages that occur in soluble rocks along preexisting discontinuities.
Joints, faults, and bedding planes sometimes occur congruently (for example, in
bedding joints). Most often, however, a distinction is made in the field between
bedding and jointing.
Uncertainty arises in many ways when dealing with discontinuities in a rock

mass. All characteristics of discontinuities, including spacing, aperture, per-
sistence, planarity, amount and kind of filling, and the corresponding hydraulic
conductivity associated with all of these features, change over distance, both
laterally and with depth. Some characteristics also change with time. Stresses
imposed by the dam and reservoir deform joints and bedding planes. Filling
materials are eroded into and out of joints under changing water pressures and
flow velocities. There is uncertainty in extrapolating measurements and descrip-
tions made at one location to the entire foundation rock mass, particularly if obser-
vations are made in boreholes, which are essentially one-dimensional. It may also
be difficult to confidently extrapolate measurements into the future not knowing
precisely what stresses and deformations will occur and what effects they will
have on the properties of rock discontinuities. Much of this report, particularly
Chapters 2 and 5, addresses the effects of stresses and deformation on the predic-
tion and calculation of foundation uplift pressures.
3.5.1 Describing and measuring discontinuities in rock
For engineering purposes, descriptions of discontinuities should be

quantitative when possible, should be pertinent to engineering usage, and should
include characteristics or properties readily determined in the field. Characteristics
of discontinuities that meet these restrictions are joint spacing (joint frequency)
and bed thickness, true orientation or attitude within the rock mass (and orienta-
tion relative to nonhorizontal excavation surfaces), and condition. Discontinuity
condition encompasses surface roughness (asperity), width of opening (aperture),
degree of weathering, and type and degree of filling.
Another feature of rock discontinuities, one that leads to perhaps the most
uncertainty in evaluating flow through jointed rock, is persistence. Persistence
describes the degree to which an individual discontinuity maintains its identity
and influence throughout a rock mass or within the boundaries of an excavation or
construction site.
Spacing. Ideally, discontinuity spacing applies to the three-dimensional rock

mass. Realistically, measurements of spacing are made in the field in one or two
dimensions. Borehole core and photolog measurements of spacing are

one-dimensional (along a line), and most rock exposure measurements are two-
dimensional (in a plane). Borehole measurements are biased in favor of discon-
tinuities lying at nearly right angles to the borehole axis and against those lying
parallel to the borehole axis because more of the former intersect the borehole. A
foundation investigation should extrapolate to three-dimensional spacing by
analyzing all complementary data from boreholes, trenches, cuts, and other
exposures.
Condition. Discontinuity condition, as discussed previously, is an important

consideration in classifying rock mass quality for engineering purposes. Shear
strength along discontinuities and the stability of the rock mass are affected by the
height and strength of surface irregularities (roughness, or asperity); the strength
and thickness of joint filling, which is often clayey and considerably weaker than
the host rock; and the strength of joint wall rock. Joint aperture controls the
secondary permeability or effective porosity of a rock mass.
Orientation. The geometry or orientation of discontinuities is described by

measuring their strike and dip. The strike of a joint or bedding plane is the direc-
tion of a horizontal line within the plane (the direction of the line of intersection
of the plane with the horizontal). Strike may be measured in degrees relative to
north, such as “north 50° west,” or degrees of azimuth measured clockwise from
north, as “310°.” The dip is the angle between 0 and 90° that the plane makes
with the horizontal, measured perpendicular to the strike. Figure 3.6 illustrates the
concepts of strike and dip and joint sets and system. A joint set is a group of more
or less parallel joints. A joint system is made up of two or more joint sets with
consistent patterns. Joint orientation influences the stability of excavations in rock.
Discontinuities that dip into an excavation, or sets of joints that intersect to form a
wedge dipping out of the excavation face, may be said to be adversely oriented
with respect to the excavation. Joint orientation is perhaps less important than
aperture and filling in assessment of uplift pressures. It is the condition of flow
within joints that controls the development of uplift pressures in a rock
foundation.
Grenoble and Amadei (1990) presented data from a study of uplift pressures
in dam foundations. The study used finite element models of the dam and the
jointed rock mass to determine the sensitivity of uplift pressures to geological
factors. Their data showed that severity of uplift pressure was independent of joint
orientation. The finite element model was for a foundation cut by an orthogonal
system of joints. Each of ten joint networks was rotated through 360° in 15° incre-
ments, and the uplift pressure was calculated by the model. Figure 3.7 graphically
shows the results of the study. Grenoble and Amadei stated that although the flat,
dashed line through the data points of the graph is a representative linear fit of the
data, there is considerable scatter in the data points.
Aperture. Engineering classifications of rock mass commonly describe joint

aperture as either “tight” or “open” (see Chapter 2 of this report for a quantified
definition). Deere (1964) recommended that, in addition to these terms, the mag-
nitude of aperture be recorded for open joints. Effective, or secondary, porosity
(fracture porosity of rocks that have little or no primary, or grain, porosity) can be

E
IK
R
ST
DI
P
a.
Joint Set 1
Joint Set 2
b.
Figure 3.6. (a) Joint orientation or attitude defined by strike and dip of joint plane
(shaded). (b) Joint system consisting of joint sets 1 and 2
estimated from analysis of the volume of open joints determined from borehole
photographic, borehole television, or other borehole viewer logging.
Borehole photography investigation of jointing in the foundation of Teton

Dam defined effective porosity as the total open-joint volume divided by the
volume of the boring (Banks 1977). For the Teton Dam analysis, investigators
determined joint condition and aperture for every joint visible in the boring walls.

Figure 3.7. Results of finite element joint orientation study. Dashed line indicates uplift force calculated
assuming a linear variation in head (after Grenoble and Amadei 1990)
Joint condition was “tight” if no aperture was present, “open” if separation of the
joint walls was persistent, and “partially open” if the joint walls did not remain
separated throughout the film record. For calculation of effective joint porosity,
the volume of open joints was taken at 100 percent; the volume of partially open
joints was halved.
Discontinuity surfaces can be examined in rock excavations if the excavation

surfaces are fresh (for example, in machine-bored tunnels). Measurements of
aperture in boreholes can be made by impression-type packers, which expand
against the borehole walls and take an imprint of wall irregularities, such as open
joints. However, ISRM (1978) emphasized that measurements of the exposed
surfaces of open discontinuities may not be representative of water-conducting
potential because wall roughness may reduce flow velocities. In addition, open
discontinuities may be filled or closed at some distance from the measured
exposure. In situ permeability testing (e.g., pressure testing, pump testing, bailing,
and falling head) is a more reliable indicator of flow through apertures.
Filling. Site investigators should describe material within the walls of a dis-
continuity in terms of its thickness, relative grain size, and, if possible, its compo-
sition (mineralogy). Fillings such as calcite and gypsum, which are subject to
removal under construction stresses or by solution, may produce greater apertures

than those initially measured when they are removed. If the thickness and nature
of such a filling are recorded, the effects of subsequent widening of the aperture
can be predicted or expected.
Fillings of cohesionless materials may flow out when the rock mass is exca-
vated. Fillings of clays with a high activity number (plasticity index divided by
percent particles smaller than 0.002 mm) can undergo considerable volume
change in the presence of varying moisture conditions. In general, the more active
a clay soil, the greater will be its volume change under changing moisture condi-
tions. Low-activity or inactive clays are relatively weak and can be washed out
from the joints. Table 3.3, from Brekke and Howard (1972), describes materials
often filling joints and the potential problems associated with them. Brecciated or
gouge fillings may ravel and be easily washed out of joints, resulting in higher
hydraulic conductivities.
Table 3.3
Material Filling Discontinuities and Associated Problems (modified
from Brekke and Howard 1972
Material Filling Discontinuity Potential Problems
Swelling clay (montmorillonite, illite, attapulgite) Subject to volume change in variable moisture
conditions. May produce swelling conditions
when confined. May cause lifting of excavation
surfaces and foundations.
Inactive clay Represents weak material between discontinuity
walls, with low shear strength if thick enough.
Can be washed out, resulting in open
discontinuity.
Low-friction metamorphic minerals (chlorite, Low resistance to sliding, especially when wet.
talc, graphite, serpentine)
Crushed rock fragments or breccia; sand-like May ravel or run out of exposed discontinuity.
gouge. Permeability may be high.
Calcite, gypsum Soluble: may later produce larger apertures than
initially measured. May be weaker than wall rock.
Asperity. Roughness, or asperity, of discontinuity walls is manifested in the

presence or absence of surface irregularities and their magnitudes. Goodman
(1968) showed that roughness and other joint conditions affect peak strength and
load deformation curves of laboratory direct shear and in situ block shear tests on
discontinuities. In addition to effects on rock shear strength, asperities on a joint
surface reduce the aperture of the joint, which directly affects the effective
porosity and permeability of the joint, of direct importance to assessment of uplift
potential. As discussed in Chapter 2, roughness in a joint lowers the critical
Reynolds number, the aperture at which flow becomes nonlaminar.
3.5.2 Effects of joint condition on groundwater flow
A study by Pahl, Bräuer, and Liedtke (1995) illustrated the importance of

open discontinuities to the fluid conductivity of a rock mass. Salt tracer experi-
ments in a fractured rock mass determined the fluid conductivity between an
injection borehole and an observation borehole. Figure 3.8 compares salt tracer

Number of open fractures per meter
Fluid conductivity
Fluid conductivity
Number of open
fractures per meter
Depth in borehole
Figure 3.8. Relationship of fluid conductivity to the distribution of open fractures

in a rock mass. Conductivity and fracture frequency data are for the
same borehole (after Pahl, Bräuer, and Liedtke 1995)
fluid conductivity to the number of open fractures in one tested borehole.

Comparison of fluid conductivity and the distribution of open fractures within
logged boreholes showed a direct relationship between fluid conductivity and the
number and distribution of open fractures. Fluid (hydraulic) conductivity of the
rock mass is important to the distribution and magnitude of uplift pressure, as
shown previously. Obviously, the number of open fractures, or fracture frequency,
which affects the magnitude of fluid conductivity, is an important attribute of rock
foundations in uplift prediction. Chapter 5 of this report analyzes the relationship
between equivalent hydraulic conductivities and distribution of joints in pressure-
tested zones in boreholes of the foundation of Libby Dam.
The length of joints and the degree to which joints are interconnected influ-
ence the distribution of uplift pressures beneath a structure. Stone and Webster
Engineering Corporation (1992) discussed the effects of joint length, degree of
interconnectivity, and aperture on uplift pressure distribution. Figure 3.9a shows
the distribution of uplift pressure along a smooth joint of constant aperture and
length L. Figure 3.9b shows the distribution of pressure along a longer joint of
length L + ∆L. The pressure at the left end of each joint is P1 and at the right end
is P2. The pressure measured at a point, a, in the shorter joint is Pa. The pressure
measured at point a in the longer joint is Pa. Pressure in each joint decreases
linearly from P1 to P2, but because the pressure decreases more slowly in the long
joint than in the short joint (the slope of the line is lower), the pressure at point a
is higher in the long joint.

Pa > Pa (from Figure 3.9 a)
Figure 3.9. Influence of joint length on uplift pressure distribution (after Stone and Webster Engineering
Corporation 1992)
Figure 3.10, also from Stone and Webster Engineering Corporation (1992),
illustrates the effects of joint interconnectivity and relative aperture on uplift pres-
sure distribution. In Figure 3.10a, seepage in the foundation is through a set of
two joints of different apertures. Most of the pressure dissipation is within the
small aperture because of frictional losses, and the pressure distribution is as
shown. In Figure 3.10b, the reservoir feeds joints that are not connected to the
tailwater of the dam. Water cannot escape through these joints and so the full,
undissipated headwater pressure exists over the entire joint area, as shown by the
horizontal portion of the dashed line. In a highly interconnected network of joints,
the foundation behaves like a porous medium, with diffuse flow, and the pressure
distribution approaches a straight line from headwater to tailwater.

Figure 3.10. Influence of relative joint aperture and joint interconnectivity on uplift
pressure distribution (after Stone and Webster Engineering
Corporation 1992)

3.5.3 The problem with faults
Deere (1981) discussed the critical role of faults in the design and stability of
concrete dam foundations. Faults are similar to shear zones in metamorphic rocks
but with additional troublesome characteristics. Unlike shear zones, faults may
occur at any orientation, not simply parallel to bedding and foliation. The fault
zone may range from centimeters to tens of meters in thickness, but it is typically
1 to 3 m thick. Because of extensive movement along the fault, the zone usually
contains broken and slickensided1 rock, fault breccia, and clayey fault gouge.
These materials act as zones of either higher or lower hydraulic conductivity. The
broken rock on either side of the fault has higher permeability. If the fault zone is
oriented so that its upstream end has access to the reservoir and its lower end
outcrops below the dam, the fault is a conduit to seepage and can cause leaking
and piping of fines. Faults oriented other than parallel to the flow may cause
blockage of flow because of the clayey gouge within the fault zone. High piezo-
metric levels, and high uplift pressures, may exist upstream of the fault, and cut
off from downstream drains. The high hydraulic gradients upstream of the fault
may cause blowouts of fine-grained gouge or soil into open joints or into drainage
holes. To accommodate fault zones, deep excavation and concrete backfilling may
be necessary. Deere stresses that the engineering geologist should take all steps in
the initial site investigation to discover and map fault zones prior to the start of
construction. Methods would include the use of aerial and satellite imagery to
locate lineations that might signify the presence of faults, mapping of depressions,
springs, geologic contacts and geomorphological (landform) indications, and
exploratory borings, trenches, adits, test pits, and shafts to characterize the faults
adequately.
An example of the impact of faults on dam foundation design and construc-

tion activities was provided by the construction of the third dam and power plant
for Grand Coulee Dam in central Washington. The forebay dam for the power
plant is a gravity-type structure with a maximum height of 200 ft above the foun-
dation. The forebay dam was constructed on the right bank of the Columbia River
downstream of the main Grand Coulee Dam. Bock, Harber, and Arai (1974) dis-
cussed problems encountered in the initial stages of construction of the forebay
dam. The foundation consists of fine- and coarse-grained granite. Geologic
investigations of the forebay dam foundation revealed two major faults. One ran
parallel to the dam axis and dipped about 60° downstream, with an apparent width
of about 10 ft. Officials decided to treat the fault by applying dental concrete to a
depth of 25 ft. As excavation progressed, however, it was apparent that the fault
was considerably wider near the center of the forebay dam. It also contained a
distinct zone of gouge on each side, separated by zones of severely jointed rock.
The entire 50-ft width of the fault had to be excavated and backfilled to a depth of
25 ft. In addition, a set of nearly horizontal drains was drilled upstream at an angle
from the toe to intercept the fault. The drains were designed to reduce the pressure
against the gouge material and to reduce foundation uplift pressures. It also
became apparent that the fault gouge was susceptible to piping. To prevent plug-
ging of the drains by eroded gouge, slotted polyvinyl chloride (PVC) pipe was
1
Slickensides are polished and striated surfaces on a fault plane. They are generally aligned
parallel to the direction of movement along the fault.

installed in the drain holes where they crossed the fault. This is an example of the
variety of problems that faults can cause in evaluating and treating a foundation.
3.6 Foundation Investigation Methods

Engineer Manual (EM) 1110-1-2908 (Headquarters, U.S. Army Corps of
Engineers, 1994) explains the concept of a guided approach and stepped proce-
dures for the design of rock foundations. Any design that involves rock masses
requires a decision-making process in which information and data must be
obtained, considered, and addressed before decisions and judgments can be made.
A coordinated team of geotechnical and structural engineers and engineering
geologists is required to ensure that rock foundation conditions and design are
properly integrated into the overall design of the structure and that the compiled
final design of the structure is safe, efficient, and economical. Investigations
address two usual analytical concerns in the design of dam foundations, bearing
capacity and sliding stability. Data that should be obtained in foundation investi-
gations during the design phase include characteristics of discontinuities, depth of
overburden, groundwater conditions, depth and degree of weathering, lithology,
physical and engineering properties of the rock mass, and loading conditions. The
analyses of rock foundations must include an evaluation of the effects of seepage
and uplift forces and of the grouting performed to reduce seepage.
Methods of determining the geological and engineering characteristics of a

dam foundation include surface geologic mapping, geophysical exploration,
boring and sampling, borehole testing, exploratory excavations, in situ testing,
laboratory testing, and groundwater pumping and pressure testing. Detailed pro-
visions for foundation investigating procedures, techniques, and methods are
provided in EM 1110-1-2908, Rock Foundations (Headquarters, U.S. Army Corps
of Engineers, 1994); EM 1110-1804, Geotechnical Investigations (Headquarters,
U.S. Army Corps of Engineers, 2001); EM 1110-1-1802, Geophysical Explora-
tion for Engineering and Environmental Investigations (Headquarters, U.S. Army
Corps of Engineers, 1995); and EM 1110-2-1901, Seepage Analysis and Control
for Dams (Headquarters, U.S. Army Corps of Engineers, 1993). This section
summarizes methods commonly used to characterize the rock foundations of large
dams.
3.6.1 Surface mapping
Geologic mapping of the damsite actually begins long before the start of
foundation investigations. Siting of the dam requires extensive areal mapping to
develop an accurate picture of the geologic framework of the area for site selec-
tion decisions. The amount of areal mapping required depends on the complexity
of the regional geology, the size of the project, and the extent to which conditions
have previously been described and mapped in existing published geologic
reports. Large-scale, detailed geologic maps are prepared for specific sites,
particularly the foundation, abutments, and appurtenant areas of the dam. The
foundation map is made during excavation and construction. It is a geologic map
with details on structural, lithologic, and hydrologic features. It can represent

structure foundation, cut slopes, and geologic features in tunnels or large cham-
bers. Mapping usually is performed after the foundation has been cleaned just
before the placement of concrete or backfill. Surface cleanup at that time usually
is sufficient to allow the recording of all geologic details on the foundation sur-
face. The engineering geologist should be familiar with the design memoranda
and should discuss the design with design engineers. Foundation geology should
be compared with the regional geologic model developed through initial project
mapping to determine if there are differences in geologic conditions that require
evaluation.
The engineering geologist should look for and map indications of adverse
conditions in the foundation rock mass. Adverse conditions affect the stability of
cut slopes, foundation settlement and bearing capacity, sliding stability of struc-
tures, and water control measures such as grouting, seepage control, and control of
uplift forces. Adverse conditions in rock include zones of weathering, soft inter-
beds in sedimentary and volcanic rocks, lateral changes in rock types and rock
properties, presence of materials subject to volume change, adversely oriented
discontinuities, highly fractured zones, faults, joints, and shear planes filled with
soft or low-resistance materials, and exceptionally hard layers that slow excava-
tion or drilling. Adverse groundwater conditions include high pore pressures and
uplift pressures, swelling materials, slaking, and piping.
The field-mapping geologist should start with a geologic interpretation, or

conceptual model, of the site, and then refine it as geologic mapping progresses.
The scale of the foundation geologic map typically ranges from 1 in. = 5 ft for
hard rock with many discontinuities to 1 in. = 50 ft for softer sedimentary or less
jointed rock. The field base map should have lines of reference for location pur-
poses. Where slopes are nearly vertical or steep and access is difficult, mapping
on large-scale photographs may be desirable. Geologic features to be included on
the map may include rock type and contacts between rock types; rock structure,
including orientation and spacing of bedding planes, joints, faults, shear zones and
other discontinuities; the shape of rock blocks (blocky or massive, tabular); and
the presence of solution cavities or voids and the materials filling them.
3.6.2 Geophysical explorations
Geophysical exploration consists of making indirect measurements from the

earth’s surface or in boreholes to obtain information about the subsurface. Geo-
physical measurements, when interpreted, provide information about the geology,
structure, and groundwater conditions in the subsurface. Geophysical explorations
are most valuable when performed early in the exploration program. They are
appropriate for rapid location and correlation of geologic features and the in situ
measurement of rock mass elastic moduli and density. The six major geophysical
exploration categories are reflection and refraction seismic, electrical resistivity,
sonic, magnetic, radar, and gravity. Geophysical explorations are usually
calibrated with a limited number of borings.

3.6.3 Boring and sampling
Subsurface investigations are typically invasive and require disturbance of the

ground to varying degrees. Most are relatively expensive and should be carefully
planned to yield the maximum amount of information. Procedures should be
followed carefully to ensure the production of high-quality data. The primary and
most versatile means of investigating the interior of the foundation, the sub-
surface, is the use of borings. Borings are used to define geologic stratigraphy and
structure, obtain samples for testing, obtain groundwater data, perform in situ
tests, obtain samples to determine engineering properties, and install instrumenta-
tion. In rock, borings are classified broadly as rock bit and core.
Rock borings not requiring samples advance using solid bits, including fish-
tail, drag bits, tri-cone and roller rock bits, or diamond plug bits. Rotary-cored
rock samples are commonly retrieved in 5- to 10-ft lengths in hollow-core barrels
equipped with diamond- or carbide-impregnated bits. Core hole diameters range
from about 1.2 to about 7.75 in. The most common hole size used by the Corps
for geotechnical investigations is NX size, with a hole diameter of about 3 in. The
use of wireline drilling, whereby the core barrel is retrieved through the drill rod
string, eliminates the need to remove the drill rods for sampling and saves a great
deal of time in deep borings. Table 3.4 lists popular core bit core and hole diame-
ters. Core recovery in zones of weak or intensely fractured rock is particularly
important because these zones are typically the critical areas in foundation loading
and stability considerations. The use of larger diameter core bits ranging from 4 to
6 in. in diameter are frequently required to produce good core in highly fractured
rock. Larger diameter core samples are also desirable for rock strength tests,
especially for testing discontinuities.
Table 3.4
Popular Diamond Core Bit Core and Hole Diameters
Bit Designation Core Diameter, in. (mm) Reaming Shell (Hole) Diameter, in. (mm)
“W” Group, “G” and “M” design
EWG (EWX), EWM 0.845 (21.5) 1.485 (37.7)
AWG (AWX), AWM 1.185 (30.0) 1.890 (48.0)
BWG (BWX), BWM 1.655 (42.0) 2.360 (59.9)
NWG (NWX), NWM 2.155 (54.7) 2.980 (75.7)
HWG 3.000 (76.2) 3.907 (99.2)
Large-diameter
2-3/4 X 3-7/8 2.690 (68.3) 3.875 (98.4)
4 X 5-1/2 3.970 (100.8) 5.495 (139.6)
6 X 7-3/4 5.970 (151.6) 7.750 (196.8)
Wireline
AQ 1-1/16 (27.0) 1-57/64 (48.0)
BQ 1-7/16 (36.5) 2-23/64 (60.0)
NQ 1-7/8 (47.6) 2-63/64 (75.8)
HQ 2-1/2 (63.5) 3-25/32 (96.0)
PQ 3-11/32 (85.0) 4-53/64 (122.6)

Most rock core borings are drilled vertically. However, inclined and horizon-
tal borings may be required to define stratification, jointing, and other discontinui-
ties adequately, or to install drain or grout holes in critical areas. Oriented core
drilling may be required if precise geological structure (orientation of disconti-
nuity planes in the subsurface) is to be evaluated from core samples. In these
procedures, the core is scribed or engraved with a special drilling tool so that its
orientation is preserved. Borehole viewing equipment, such as cameras, can also
be used to determine in situ orientation.
There is always a degree of uncertainty in interpreting subsurface geologic

structure from small-diameter boreholes. Core breakage from drill action may
mask natural joints and bedding planes. Orientation of discontinuities is obtained
somewhat indirectly and may be disturbed during the handling of samples. The
persistence of discontinuities within the rock mass is particularly hard to gauge
from a borehole. The properties, particularly strength, of cored samples may
overestimate the strength of rock mass, because the effects of planes of weakness
may be missed in a small sample. Large-diameter borings, or calyx holes, 2 ft or
more in diameter, are sometimes used in large or critical structures. Calyx holes
permit direct examination of the borehole walls and provide access for performing
in situ tests and obtaining high-quality undisturbed samples.
Core logging is usually performed immediately after the core is retrieved from
the boring, while natural discontinuities are fresh and the rock has not been
exposed to deterioration from stress relief and changing moisture conditions. The
core log commonly includes the rock type designation and the name of the geo-
logic unit, if known. The core log provides a field determination of the relative
strength of the rock, the degree of weathering, the texture, the structure, and the
presence and condition of discontinuities. The latter may include orientation with
respect to the core axis, surface roughness, nature of infilling or coating, presence
of staining, and tightness or aperture. Other features of the cored rock include
color; swelling and slaking properties, if appropriate; inclusions, such as fossils or
minerals; and the presence of solution cavities or other voids. Rock quality
indexes, such as rock quality designation (RQD), may also be determined during
the core logging process.
3.6.4 Borehole examination and testing
A wide array of downhole geophysical probes is available to measure geologi-

cal properties and to supplement core sampling. Borehole electrical probes mea-
sure conductivity or resistivity of fluids within the formation and aid in correlating
strata between boreholes. Downhole radiation tools measure natural or emitted
radiation of formation fluids as an aid to correlation of strata and for determina-
tion of type of rock, rock density, and formation hydraulic conductivity. Borehole
imagery devices delineate voids and discontinuities in the wall of the borehole
using acoustic energy. Mechanical and acoustic calipers measure borehole
diameter.

3.6.5 Exploratory excavations (tunnels, shafts, drifts, test pits,
trenches)
The complexity, extent, and size of some shear zones and faults in a dam
foundation rock mass require a larger surface exposure for adequate examination
and testing than is offered by borings. In such cases, the excavation of test pits,
trenches, or drifts may be necessary to evaluate the feature. Test pits and trenches
can be constructed quickly and economically by bulldozers, backhoes, pans, drag-
lines, or ditching machines in rippable rock and soil. Many test pits and most
exploratory tunnels in rock require drilling and blasting and are relatively expen-
sive. Test pits and trenches generally are used only above the groundwater level.
Exploratory trench excavations are often used in fault evaluation studies. Large-
diameter calyx holes have been used successfully on some jobs to provide access
for direct observation of critical features in the foundations.
Exploratory tunnels, or drifts, permit detailed examination of the condition

and orientation of rock structures. Commonly used in the foundations and abut-
ments of large dams, they are particularly appropriate in defining the extent of
marginal strength rock or adverse rock structure detected in surface mapping and
borings. For large projects where high loads will be transmitted to the foundation,
tunnels and large shafts provide the only practical means for testing in-place rock
at locations and in directions corresponding to the structural loading. The geologic
information gained from careful mapping of the tunnel provides a critical supple-
ment to interpretations based on data from surface mapping and other sources.
Zones of excessive seepage or blockage of groundwater flow in buried shear or
fault zones may have a strong effect on the development of uplift pressures
following dam construction. Exploratory tunnels offer a means of detecting and
examining such zones directly. Exploratory tunnels often serve a multiple purpose
by providing access for drainage and grouting holes, for post-construction obser-
vations, and for utility conduits.
3.6.6 In situ testing
In situ tests are often the best means for determining the engineering proper-
ties of subsurface materials and, in some cases, may be the only way to obtain
meaningful results. In situ rock tests are performed to determine in situ stresses
and deformation properties (elastic moduli) of the jointed rock mass, shear
strength of jointed rock masses or critically weak zones, such as shear zones,
within the rock mass, and residual stresses along discontinuities or weak zones.
Table 3.5 summarizes the types and purposes of in situ tests in rock and soil.
Pressure tests, which measure the hydraulic properties of the rock mass or of
limited zones of rock, are most applicable to investigations of the potential for
uplift pressures. Pressure testing in foundations of gravity dams is discussed at
length in Chapters 2 and 5 of this report.

Table 3.5
In Situ Tests for Rock and Soil (Headquarters, U.S. Army Corps of
Engineers 2001)
Applicability to
Purpose of Test Type of Test Soil Rock
Shear strength Plate bearing or jacking X X1
2
Pressuremeter X
Uniaxial compressive2 X
2
Borehole jacking X
Bearing capacity Plate bearing X X1
Stress conditions Hydraulic fracturing X X
Pressuremeter X X1
Overcoring X
Flatjack X
Uniaxial (tunnel) jacking X X
Borehole jacking2 X
Chamber (gallery) pressure2 X
Mass deformability Geophysical (refraction) X X
Pressuremeter or dilatometer X X1
Plate bearing X X
Uniaxial (tunnel) jacking X X
Borehole jacking2 X
2
Chamber (gallery) pressure X
1
Primarily for clay shales, badly decomposed, or moderately soft rocks, and rock with soft seams.
2
Less frequently used.
3.6.7 Laboratory testing
Laboratory tests provide data on physical and hydrological properties of

natural materials, determine index values for identification and correlation by
means of classification tests, and define the engineering properties in parameters
usable for foundation design. A list of references for laboratory tests pertinent to
foundation investigations in rock is provided in Table 3.6.
3.6.8 Groundwater and foundation seepage investigations
Groundwater investigations for dam foundations in rock provide baseline data

on water table and piezometric levels in the vicinity of the dam, areas or zones of
pore pressure and potential uplift pressure, and estimates of seepage and the loca-
tion of sources of seepage. Groundwater studies include observation and measure-
ment of flows from springs and seeps and water levels in existing production
wells, exploratory boreholes, observation wells, and piezometers. Foundation
investigations include field and laboratory tests for permeability and hydraulic
conductivity, pressure testing of rock strata in boreholes, and pumping tests for
determination of rock mass hydraulic conductivity. This information is used with
site and regional geologic information to determine water table or piezometric
elevations, fluctuations in groundwater elevations, direction and rate of seepage
flow in the foundation area, and potential for leakage beneath the dam. The most

Table 3.6
Laboratory Classification and Index Tests for Rock (after
Headquarters, U.S. Army Corps of Engineers 2001)
Test Test Method Remarks
Unconfined (uniaxial) RTH1 111 Primary index test for strength and deformability of
compression intact rock
Water content RTH 106 Indirect indication of porosity of rock or clay content
of sedimentary rock
Pulse velocities and elastic RTH 110 Index of compressional wave velocity and elastic
constants constants for correlation with in situ geophysical
test results
Rebound number RTH 105 Index of relative hardness of intact rock cores
Permeability (hydraulic RTH 114 Intact rock (no joints or other major discontinuities)
conductivity)
Petrographic examination RTH 102 Performed on representative cores of each
significant lithologic unit
Specific gravity and RTH 107 Indirect indication of soundness and deformability
absorption
Unit weight and total RTH 109 Indirect indication of weathering and soundness
porosity
Point load testing (also RTH 325 Used to predict other strength parameters with
performed in field) which it is correlated
Elastic moduli from uniaxial RTH 201 Intact rock cores
compression test
Triaxial compressive RTH 202 Deformation and shear strength of core containing
strength inclined discontinuities
Direct shear strength RTH 203 Strength along planes of weakness or rock-
concrete contact
1
Rock Testing Handbook (Geotechnical Laboratory 1993).
reliable means for determining water levels and monitoring pore pressures is the
use of piezometers and observation wells. Piezometers measure pore pressures at a
point in the subsurface by means of a porous tip, sealed at a particular depth, and
a standpipe or electrical connection to the surface for recording the pressure.
Observation wells commonly have a section of open hole or slotted or perforated
pipe, 2 or more feet in length, connected to the surface by a standpipe. Observa-
tion wells usually measure a composite water level over a considerable length of
the aquifer. Table 3.7 describes instruments for measuring piezometric pressure.
Piezometers and observation wells allow measurement of fluctuations in

piezometric levels over time. Periodic readings are commonly taken to monitor
pressure changes that may develop with reservoir fluctuation and increase or
decrease in seepage flows. Locations of piezometers should be selected to provide
the data necessary to detect and monitor pore pressure levels and changes and to
aid in design of seepage and uplift control measures, such as the installation of
drains and grout curtains. The presence of a confined zone or several zones each
with a different water level requires the use of piezometers to confine and separate
each level. Placing piezometers within key open discontinuities, such as open
bedding or shear zones, and monitoring pressures periodically would permit
indirect observation and confirmation of deformation within the foundation by
reservoir stresses.

Table 3.7
Instruments for Measuring Piezometric Pressure (Headquarters, U.S. Army Corps of
Engineers 1995a)
Instrument Type Advantages Limitations1
Observation well Easy installation Provides vertical connection between
Field readable strata and should only be used in
continuously permeable strata
Open standpipe piezometer Reliable Time lag can be a factor
Long successful performance record Subject to damage by construction
Self-de-airing if inside diameter of equipment and by vertical
standpipe is adequate compression of soil around standpipe
Integrity of seal can be checked after Extension of standpipe through
installation embankment fill interrupts
Can be used to determine permeability construction and may cause inferior
Readings can be made by installing compaction
pressure transducer or sonic sounder Possible freezing problems
in standpipe Porous filter can plug owing to repeated
water inflow and outflow
Twin-tube hydraulic piezometer Buried components have no moving Application generally limited to long-term
parts monitoring of pore water pressure in
Reliable when maintained embankment dams
Long successful performance record Elaborate terminal arrangements needed
When installed in fill, integrity can be Tubing must not be significantly above
checked after installation minimum piezometric elevation
Piezometer cavity can be flushed Periodic flushing is required
Can be used to determine permeability Possible freezing problems
Short time lag Attention to many details is necessary
Can be used to read negative pore water
pressures
Pneumatic piezometer (Embedded) Short time lag Requires a gas supply
Calibrated part of system accessible Installation, calibration, and maintenance
Minimum interference to construction; require care
level of tubes and readout
independent of level of tip
No freezing problems
Vibrating wire piezometer (Embedded) Easy to read Potential for zero drift (Special
Short time lag manufacturing techniques required to
Minimum interference to construction; minimize zero drift2)
level of lead wires and readout Need for lightning protection should be
independent of level of tip evaluated
Lead wire effects minimal
Can be used to read negative pore water
pressures
Electrical resistance piezometer Easy to read Potential lead wire effects unless
(Embedded) Short time lag converted to 4 to 20 milliamps
Minimum interference to construction; Errors caused by moisture and corrosion
level of lead wires and readout are possible
independent of level of tip Need for lightning protection should be
Can be used to read negative pore water evaluated
pressures
1
Diaphragm piezometer readings indicate the head above the piezometer, and the elevation of the piezometer must be
measured or estimated if piezometric elevation is required. All diaphragm piezometers, except those provided with a vent to the
atmosphere, are sensitive to barometric pressure changes. If piezometer pipes, tubes, or cables are carried up through fill, there
will be significant interruption to construction and the probability of inferior compaction.
2
See Dunnicliff (1988).
Source: Dunnicliff (1988).

Hydraulic conductivity of the rock mass or of porous strata can be calculated
from pumping test data, slug (single-hole falling head or constant head) tests, or
pressure tests. Pumping tests employ a combination of a pumping, or discharge,
well and observation wells at variable spacing out from the well. The flow rate of
the well and the measured drawdown in the observation wells supply the neces-
sary data to calculate the average hydraulic conductivity of the rock within the
zone of influence of the pumping well. A problem with pumping test data in rock
in which flow occurs anisotropically within discontinuities, rather than through a
pervious rock mass, is that the influence of one or more conductive discontinuities
can be overlooked. Pressure tests of specific zones of rock are more useful in
delineating zones of hydraulic conductivity in jointed rock. In addition, pressure
tests in intervals more accurately define the depth of permeable rock, because rock
mass conductivity normally decreases with depth. This kind of information is
crucial to designing drainage and grouting curtains to control seepage and uplift in
dam foundations.

4 Selection of Case History
for Uplift Investigation
This chapter documents the process of selection of a dam for case-history

study in the uplift uncertainty investigation. The authors reviewed historical docu-
ments and data available for seven U.S. Army Corps of Engineers concrete or
concrete and earth-fill dams nationwide. Dams with well-documented foundation
geological investigations, particularly with respect to groundwater and joint
conditions, and abundant and well-documented uplift pressure monitoring data
had the highest potential for selection as a case history.
On recommendation of the Office of the Chief of Engineers, selection of the

potential pilot study candidates focused on two geographic areas: the northwestern
United States and the Tennessee–Kentucky area, where dams were instrumented
by uplift gauges during construction. Dams under consideration were in four
Corps Districts: Seattle, Portland, Walla Walla, and Nashville. The dams under
consideration spanned a period of 25 to 55 years from completion of construction
and were all continuously monitored by instruments for uplift.
Two types of dams were considered: entirely monolithic concrete gravity

dams in narrow rocky valleys of the Northwest, and much lower dams consisting
of earth-fill-compacted embankments with concrete gravity dams in the center.
The latter types were all in the Nashville District.
The following aspects were weighed in consideration of potential candidates

for the uplift uncertainty study:
a. Availability of reliable structural and geologic documentation for each

dam.
b. Availability and suitability of subsurface exploration performed before,
during, and after construction of each dam.
c. Location of instruments in monoliths in proximity to available geologic
data.
d. Performance of instrumentation as recorded and published in Periodic
Inspection Reports for each dam.
The dams are described in the following paragraphs, with the most suitable
dams described first.
56 Chapter 4 Selection of Case History for Uplift Investigation

4.1 Libby Dam
4.1.1 Description
Libby Dam is located in northwest Montana on the Kootenai River 17 miles

upstream of Libby, Montana, and 221.9 miles from the confluence of the
Kootenai with the Columbia River. Initial geological investigations were con-
ducted between 1951 and 1954. The foundation investigation was completed in
1965, but the final foundation report was not published until 1979. Construction
began in 1967 and was completed in July 1973. Figure 4.1 presents a plan and
upstream view of the dam.
Libby is a concrete gravity structure, 420 ft high from top of rock and 370 ft
above the streambed at its highest monolith. It is 350 ft high from the streambed at
monolith 23, where the width of the base is 250 ft. Maximum pool elevation of
2,459 ft mean sea level (MSL) was not reached until the summer of 1974 because
of operational and safety problems during construction and lack of flow in the
river during the period 1972-1973. The minimum regulated pool elevation is
2,287 ft. The maximum length of the crest is 3,033 ft. The tailwater elevation at
Libby Dam varies between 2,198.9 ft and 2,117.3 ft with an average, determined
on a frequency basis, of 2,124.7 ft.
4.1.2 Foundation geology
An initial investigation started in 1961 prior to construction. A total of 225

exploration holes were drilled in the vicinity of the preliminary and final design
dam axis. Three test pits excavated in the extensive overburden section, near the
right abutment, permitted better sampling of the thick overburden section in this
area and more realistic design assumptions for the cut slope in the right abutment.
Four adits were excavated in the abutment areas. One was excavated into rock
adjacent to the right abutment. It was 208.3 ft in length at elevation 2,261.7 ft, and
the portal was a short distance from the toe of the dam. Adit No. 2 was 201 ft, and
adits No. 3 and No. 4 were 86.5 ft and 85.5 ft long, respectively. Adit No. 1 was
left open to provide inspection for possible water seepage on the right abutment.
Adit No. 4 was incorporated in the dam as a portion of the downstream drainage
gallery. The remaining adits were backfilled with concrete.
Contract specifications called for foundation exploration to confirm the design

evaluation of the foundation during construction. Exploration consisted of a mini-
mum of one NX core hole drilled 40 ft into the foundation rock for each abutment
monolith founded above elevation 2,300 ft MSL, and 60 ft into the foundation
rock for each abutment monolith founded below elevation 2,300 ft MSL. In addi-
tion, eight air track probe holes were drilled to establish the extent of the rock-
surface low at the downstream toes of the dam in the vicinity of monoliths 17 and
18. Drill action and cuttings were monitored to establish top of rock. Geologic
logging of foundation exploration borings for Libby Dam provided more detailed
information on discontinuities than that for Dworshak Dam, discussed in the next
Chapter 4 Selection of Case History for Uplift Investigation 57

Figure 4.1. Libby Dam, plan and view looking upstream

section. In the typical Libby Dam foundation boring log, the column labeled
“Graphic Log” apparently identifies every visible joint in the core. Major joints
and other discontinuities are described under “Description of Materials” and in the
“Remarks” columns of the boring logs.
Rock at the damsite, described in the foundation report (U.S. Army Engineer
District, Seattle 1979) consists mainly of quartzite, metasandstone, and siliceous
argillite (a weakly metamorphosed mudstone). The foundation has a well-
developed fracture system consisting of several kinds and sets of joints. Bedding
joints strike 330o (azimuth) and dip 40o to 45 o west. A prominent set of east-west
striking, high-angle shear joints dips 60o to 80o north or south. Where shear joints
intersect bedding joints, wedges were formed in the valley walls. Generally north-
south striking “relaxation joints” dip 50o to 80o east. “Tension” joints strike north-
east and dip at moderate to high angles to the southeast. Other tension joints strike
parallel to bedding and dip at right angles to bedding, and are probably related to
folding. Many low-angle rebound joints (formed during unloading of the rock
mass) have random strikes. Some prominent bedding joints are filled with gouge
and are slickensided, evidence of movement, and are considered faults. Several
episodes of movement were noted along the faults. Intersection of faults with east-
west and north-south trending joints has broken the rock mass into discrete
blocks. Certain joints were open to a considerable depth, possibly caused by
unloading after glaciation.
The grout curtain is composed of three zones: a tertiary zone 40 ft into rock, a
secondary zone 90 ft into rock, and a primary zone 160 ft into rock. In the valley
section of the dam (monoliths 18 through 27), the grout holes were inclined 25°
from vertical upstream and 15° from vertical toward the left abutment. Grouting
holes are on 5-ft centers in monolith 23. Figure 4.2 is a view upstream at the
section through the grout drainage gallery along the axis of the dam. Figure 4.3 is
a section through monoliths 22, 23, and 24 showing the arrangement of grouting
holes.
4.1.3 Instrumentation
Two galleries at the lower section of the dam provide access to uplift gauges
and tops of drain holes. The galleries run parallel to the axis of the dam and are
approximately 100 ft apart and at varying elevations. At monolith 23, the galleries
are at elevations 2,078.52 ft and 2,080 ft, respectively. The drains in the two
galleries in monolith 23 are between 60 and 150 ft in length. Drain holes were
drilled on 10-ft centers through 1/2-in. I.D. galvanized pipes, 5 ft long, embedded
in the concrete. All drain holes were 3 in. in diameter and were drilled in a plane
parallel to the dam axis. In monoliths 18 through 38, all drain holes were drilled at
an angle of 15° off vertical towards the left abutment.
Uplift pressure cells were installed under six monoliths: two abutment mono-
liths (14 and 41) and four valley monoliths (18, 23, 29, and 34). Uplift block
sensors are in monoliths 18, 19, 23, 34, and 41. Seattle District provided a com-
plete set of gauge readings for these monoliths. The data were also graphically
represented in the latest inspection report published by the District. The records

Figure 4.2. Libby Dam, upstream gallery longitudinal section

63' 63'
17+08
17+71
16+45
Monolith Monolith Monolith
22 23 24
Drains Grout
@ 10' center
EL. 2080.0 @ 5' center
EL. 2077.5 +
_ 3'
5' Concrete
Rock
15°
Drain
Grout
0 10 20 30 40 50 60 70 80 90 100
Scale in Feet
Figure 4.3. View upstream of part of upstream grout/drainage gallery, Libby Dam

span a period from January 1981 to July 1999. Monolith 23 was selected for
further review because it has one of the deepest forebays. To illustrate further the
typical behavioral pattern of the gauges, Figure 4.4 shows the relationship
between forebay elevation and total head recorded on gauge P23R1 during 1989.
There is an almost linear relationship in upward and downward trend during
forebay filling and emptying.
Figure 4.5 combines the readings of Gauge P23C1 for a period of 6 years.
There is no longer a single, flat, rising loop but a multitude of pathways. The
general trend is still linear. The plots are for monolith 23, first row of gauges,
located downstream near the forebay.
Figure 4.4. Total head recorded on Gauge P23R1 versus forebay reading, 1989,
Libby Dam
4.1.4 Assessment of data quality
Available Periodic Inspection Reports were studied to identify anomalies

occurring during uplift monitoring periods. The gauges were read in all monoliths
where uplift pressure gauges were installed: monoliths 14, 18, 23, 29, 34, and 41.
The uplift gauges were read on a regular monthly basis to monitor trends in uplift
pressure and the effectiveness of the drain system. During all recording periods,

Figure 4.5. Readings for 6 years from Gauge P23C1, Libby Dam
the uplift pressure gradients downstream of the grout curtain were far below the
design assumptions. Some of the uplift pressures upstream of the grout curtain
were above design assumptions, but the total effect of the actual uplift was below
that assumed for maximum in design. Data for the recorded period were con-
sistent. Some uplift pressure cells exhibited upward trends, but further investi-
gation showed the cause to be air trapped in the gauges. When bleeder valves
were installed to correct inaccurate readings, the gauges presented actual uplift
pressures, expressed as total head.
Total head includes the hydrostatic pressure combined with elevation pres-
sure. A detailed review and preliminary evaluation of the data from the gauges at
various monoliths indicated that monolith 23 had consistently reliable readings
with respect to forebay fluctuation. A major criterion for the selection of a mono-
lith to be used in the uplift uncertainty study was that the proximity of the gauges
to the forebay was sufficient to show the effect of the forebay fluctuation.
Readings from gauges located farther from the forebay, especially those beyond
the drain holes, did not offer a clear interpretation of the relationship between the
forebay and gauge reading. The drain effectiveness, maintenance records, and
resulting variable flow could introduce additional variables and thus obscure the
nature of the seepage uplift phenomenon. Data for monolith 23 were carefully
reviewed.

Plots of the annual cycle of forebay readings versus gauge readings showed
that year 1990 would best merit further study. Gauge readings recorded for the
latest available annual cycle should serve as a basis for further assessment of the
numerical flow modeling procedure currently under development at the
U.S. Army Engineer Research and Development Center (ERDC). Geologic and
instrumentation data for Libby Dam were rated “good.”
4.2 Dworshak Dam

4.2.1 Description
Dworshak Dam, originally named Bruces Eddy Dam, is a straight concrete

gravity dam located in a narrow canyon of the North Fork Clearwater River,
42 miles east of Lewiston, Idaho. Construction began in July 1966, and the dam
was operational for flood control by June 1972. It is 3,287 ft long and 717 ft high
at the lowest point of foundation excavation. The dam is the highest straight-axis
concrete dam in the western hemisphere, the 22nd highest dam in the world, and
the third highest in the United States. Maximum design pool elevation of the
forebay is 1,604.9 ft, crest elevation is 1,600 ft, and design pool elevation is
1,540 ft. The dam was built between 1967 and 1972. Geological exploration
preceded the construction of the dam by almost a decade. Figure 4.6 is a plan
view of the dam. The dam is oriented in almost a straight line from northwest to
southeast with water impounded to the north. It is accessible by road from the
nearby cities of Ashaka, Elk River, and Orofino, Idaho. Figure 4.7 shows the
elevation of the dam looking upstream, with a full view of all 58 monoliths. The
highest monoliths are (in this view) 23, 24, 25, and 26. Figure 4.8 is section
through the dam at station 25+40. The section shows the general structure, the
location of the drains and access galleries, the grout curtain, and contours of the
original and excavated ground surface.
The site of the dam was extensively investigated in the mid-1960s (about
1963-1964). A plan of exploration dated March 1971 is included as Figure 4.9.
The exploration consisted of NX core drill holes, AX core drill holes, calyx holes,
and several adits spread over an area of 2,400 by 1,500 ft. A total of 245 borings
were drilled in the immediate vicinity of the dam. There are five adits within a
distance of 600 ft from the centerline of the dam.
Dworshak was initially considered a good candidate for the uplift uncertainty
study. Approximately 13 borings, located within 200 ft of the dam’s longitudinal
axis, coincided with the highest portion of the dam. Borings vary in depth
between 55 and 125 ft. Additional exploration was performed as a part of under-
seepage study for the purpose of grouting in 1989 and 1996. Borings in closest
proximity to selected monolith 23 were DH-461, DH-250, DH-252 and DH-90 .

Figure 4.6. Plan view of Dworshak Dam

Figure 4.7. Elevation of Dworshak Dam, looking upstream

Figure 4.8. Section through Dworshak Dam at station 25+40
The foundation report for Dworshak Dam (U.S. Army Engineer District,
Walla Walla 1979) reported that the dam is founded on gneiss and amphibolite1 of
pre-Cambrian aged Orofino metamorphic units. Design Memorandum (DM)
No. 6 (U.S. Army Engineer District, Walla Walla, 1964) identifies the rock as a
granite gneiss (a footnote on page 3-1 of DM No. 6 states that subsequent geo-
logic investigation by the U.S. Geological Survey (Hietanen 1962) identified the
foundation rock as quartz diorite and tonalite). The foundation rock contains
discontinuities in the form of joints, shear zones, and porphyritic dikes. DM No. 6
states (p. 3-8) that the predominant joint [set] strikes 5o (azimuth) and dips 70o to
90o east. Less prominent joint sets strike 10o with dip 80o to 90o west, 350o with
dip 35o west, and 340o with dip 65o west. Joint attitudes of DM No. 6 were deter-
mined in four foundation adits excavated prior to construction of the dam. Shears
mapped in the foundation adits were in two sets, a major set striking 4o to 35o and
dipping 47o to 80o east and a minor set striking 275o to 296o and dipping 35o to
80o south.
The foundation rock is generally massive and contains no distinctive layering,

such as marker beds, to permit subdivision of the rock mass. For design purposes,
the rock was classified by the degree of weathering and the presence of shears and
1
Amphibolite is a metamorphic rock consisting of the minerals amphibole and dark-colored
plagioclase feldspar, with little or no quartz.

68
Chapter 4
Figure 4.9. Plan of exploration borings for Dworshak Dam (March 1971)
Selection of Case History for Uplift Investigation

shear zones. Weathering categories were highly weathered, moderately wea-
thered, lightly or slightly weathered, and fresh, or unweathered. Foundation
excavation requirements were determined by rock quality, which was defined by
degree of weathering. Portions of the dam higher than 150 ft were to be founded
on no worse than lightly weathered rock (showing only slight alteration of min-
erals with no kaolinization of feldspars). Portions less than 150 ft high were to be
founded on no worse than moderately weathered rock (showing only partial
kaolinization, with some volume change). There were also foundation excavation
slope restrictions.
Several shear zones up to several feet wide and persisting for up to several
hundred feet across the foundation were mapped in detail and remedied by dental
treatment. Shear zones are faults containing slickensides or crushed and clayey
gouge, evidence of movement along the faults. Gouge material was removed and
the cavities backfilled with concrete. Some of the shear zones and zones of shear
intersections were so large and extensive that drilling and blasting were required
to remove unacceptable material.
DM No. 6 (U.S. Army Engineer District, Walla Walla, 1964) evaluated the
results of 1,234 pressure tests conducted in 139 drill holes in the foundation of
Dworshak Dam. Data from water pressure tests indicated that the rock “tightened”
with depth. A scatter diagram provided in DM No. 6 (p. 5-3) was interpreted as
showing “a progressive decrease in maximum observed permeability with
depth….” The DM stated that because many pressure tests showed no water take,
the rock matrix was impermeable. By default, permeability of the rock mass was
attributed to the presence of fracturing. The scatter diagram showed higher
permeability near the top of the rock mass in lightly weathered rock, indicating a
decrease in fracture frequency with depth. Observations in the test adits showed
that water from the rock mass was insufficient to cause observable flow from the
adits, a further indication of the tightness of the rock and independent
confirmation of the pressure test interpretations.
Hydraulic tests performed on foundation rock during the exploration phase

reported a rock permeability varying between 10-5 and 10-3 ft/min, depending on
the depth from the excavated rock surface and the location. U.S. Army Engineer
District, Walla Walla (1979), offered a wealth of documentation about borings
and size and location of rock fractures.
Dworshak Dam consists of 58 monoliths. Uplift gauges are located in mono-

liths 13 through 36. A total of 65 uplift pressure measuring devices were installed
during dam construction to monitor grout curtain effectiveness and foundation
drainage system performance. The highest concentration of the gauges is in mono-
liths 21, 23, 24, 25, 28, and 31. The highest monolith, monolith 22, is not the best
instrumented one. Refer to Figure 4.10 for locations of the uplift gauges. Fig-
ure 4.11 is a detail of the centrally located monoliths 23 and 25, the best instru-
mented of all the monoliths. All gauges consist of perforated pipes drilled 4 ft into
the rock. Gauges in the second row, designated P**2, were drilled 6 ft into rock

Figure 4.10. Locations of uplift gauges, Dworshak Dam

Figure 4.11. Detail of instrumentation, monoliths 23 and 25, Dworshak Dam
because they coincided with the drains. Monolith 23 is approximately in the

middle of the dam and corresponds to the second highest forebay water elevation.
The dam has several access galleries at different elevations mutually connected for
maintenance and monitoring of instruments. Extensive instrumentation monitors
internal stresses and concrete temperature in addition to uplift pressures.

Monolith 23 was chosen for more detailed study in the uplift uncertainty proj-
ect after examination of data quality of periodic inspection reports. Nine of a total
of 65 uplift measuring gauges are located in monolith 23. The uplift pressure
monitoring gauges in monolith 23 consist of five pipe piezometers and four trans-
ducers. Three piezometers are located between the drainage gallery and forebay,
and one piezometer is located in the gallery itself. One piezometer is located
111 ft downstream.
Inspection of Dworshak Dam began in 1972 when the dam was still under
construction. No gauges were read at that time. Reporting of uplift pressures
started in 1975. Inspection Brochure No. 4, dated April 1975, recorded readings
for some gauges in monoliths 21 and 24. The brochure indicated that gauges in
the first row downstream of the forebay exceeded design uplift pressure during
periods of high pool elevations. Inspection Brochure No. 5, June 1975, confirmed
this situation. Inspection Brochure No. 7, 1981, listed only gauges for mono-
lith 21. Gauges close to the forebay reflected the fluctuation of the forebay, with
an unusually steady rising trend over the recorded time span for both peak and
bottom readings. The remaining gauges stagnated at the same reading with no
response to forebay changes. A possible explanation for this behavior is that the
readings in those gauges could be triggered only at uplift pressures that were
never reached during the recording period. Design pressures were thus grossly
overestimated, while uplift pressures next to the forebay were slightly underesti-
mated. The effectiveness of the grouting curtain and drains in connection with this
behavior suggested a strong influence downstream from the drains and grouting
curtain. Inspection Brochure No. 9, June 1989, reported on 30 gauges in mono-
liths 23, 24, and 25. Gauges in monolith 25, closer to the forebay, exceeded
design pressures during the period 1980 to 1993. The remaining gauges had read-
ings much under design pressures. Some gauges indicated pressures well under
tailwater elevation.
The authors concentrated efforts on monolith 23, which showed a greater

consistency in data over a longer period of recording than other monoliths.
Gauges P23X, P230, and P231 are located upstream of the drain holes. Gauge
P232 coincided with a drain hole, and Gauges P234, P235, P237, P238, and P239
are located downstream behind drains. Gauges P23X, P230, P240, P241, and
P242 consistently exceeded design uplift pressures for stages of the forebay above
elevation 1,600 ft. Plots relating forebay readings to gauge readings for selected
periods of time were fit to evaluate the behavioral pattern of the gauges and the
quality of the data. Figure 4.12 is one such plot. In spite of considerable scatter,
there is a distinct linearity in hysteresis of the rising and falling of the forebay in
the response for Gauge P23X. Contrary to readings of Gauge P23X, Gauge P230
(Figure 4.13) shows a distinctly two-pronged trend in the plot. The upper portion
of the plot is for the initial period after dam filling. The lower branch of the plot is
for the latter decade of recording. The significance is that the latter behavior sug-
gests a nonlinear relationship between the forebay elevation and the gauge read-
ing. For the nonlinear behavior, an appropriate explanation would be that the rock

Figure 4.12. Forebay level versus Gauge P23X reading, Dworshak Dam
fracture system underlying the dam may be opening and closing, creating non-
constant flow conditions.
Modeling of this situation would be very complicated. Figure 4.14 is a plot of

the relationship between forebay changes and gauge readings in feet of total head.
Although there is a hysteresis, it forms an oddly shaped loop.
Geologic data for Dworshak were rated “good,” and instrumentation data
were rated “sufficient.”
4.3 Green Peter Dam

4.3.1 Description
Green Peter Dam is located on the Middle Santiam River, 4.7 miles above its
confluence with the South Santiam River and approximately 30 miles southeast of
Albany, Oregon. Construction of the dam began in 1963, and the dam was dedi-
cated in June 1967. Green Peter is a 330-ft-high concrete monolith structure,
1,455 ft long at the crest and 260 ft wide at the foundation-rock interface. Design

Figure 4.13. Forebay level versus Gauge P230 reading, Dworshak Dam
maximum pool elevation is 1,015 ft. The dam is divided into 29 monoliths vary-
ing in width from 90 to 120 ft. The highest monoliths are 12 through 21, in the
center of the structure (Figure 4.15). Green Peter Dam is oriented west-northwest
to east-southeast. Water is impounded from the north. Two sections through the
dam are shown in Figure 4.16: one through the spillway and the other through the
powerhouse and penstock.
Geological exploration occurred in several stages. Most borings were

emplaced between 1959 and 1962. A total of 297 borings and calyx holes were
drilled in the immediate vicinity of the dam. Borings consist of core borings,
churn holes, and 36-in. and 42-in. calyx holes. In addition, 12 exploration drifts
and cross cuts were cut (Figure 4.17). Some drill holes were equipped with
piezometers. Approximately 40 borings drilled in the highest portion of the dam
were suitable for this study.
The dam lies in the Western Cascades geologic province and is characterized
by the presence of volcanically derived sediments, lava flows, and pyroclastic
materials with local intrusives. The foundation report (U.S. Army Engineer

Figure 4.14. Forebay level versus Gauge P230 reading, 1987, Dworshak Dam
District, Portland, 1969) stated that the foundation geology is complex, with major
lava flows separated by four interbeds of pyroclastics (Figure 3.3). A number of
shear zones, some containing clay or gouge, cut the foundation. Rocks have been
folded, jointed, and faulted by regional uplift. Groundwater occupies closely
jointed rock in semiconfined conditions. Most of the groundwater flows were
identified in open joints, where the flows were not entirely laminar. With
increased hydrostatic head, some clay materials were expected to migrate down-
gradient, thus gradually changing the flow pattern. Two drainage systems were
installed in the dam in addition to the grout curtain. Of several hundred borings
drilled for this dam, about 47 coincided with the actual dam, some exceeding a
depth of 100 ft. Figure 4-18 is a geologic section through the foundation along the
dam base line. The foundation report for Green Peter Dam (U.S. Army Engineer
District, Portland, 1969) provides much information on stratigraphy, geologic
structure, and lithology through detailed boring logs and cross sections.
A total of 35 uplift gauges were monitored at this dam in monoliths 7 through

26. Initially, Gauges 14A, PU-14-2, PU-20-2, PU-20-1, and PU-14-2A provided
useful information. The latest periodic inspection brochures available indicated

Figure 4.15. Plan of Green Peter Dam

Figure 4.16. Sections through Green Peter Dam

78
Chapter 4
Figure 4.17. Plan of Green Peter Dam showing locations of borings and drifts

Chapter 4
79
Figure 4.18. Geologic section along base line, Green Peter Dam
that six gauges may have records correlated to the fluctuation of the forebay:
Gauges 14-2, 14-1B, 18A, 18-1A, 19-1B, and 20-1. Locations of the instruments
related to uplift pressure measurement and geology are shown in plan view and
longitudinal section in Figure 4.19. Gauges marked as “U” designate uplift
gauges. Piezometers are designated as “P.” The most useful gauges for this project
would be those in monoliths 12, 16, and perhaps 21. Gauges in these high
monoliths are distributed through the foundation in a line from upstream to
downstream. Unfortunately, gauges in these monoliths did not provide much
useful information.
A substantial number of gauges clustered together is necessary to evaluate the

behavioral pattern of subsurface seepage flow that is inducing uplift. One or two
gauges in a monolith are not sufficient for a reliable model of the uplift pattern.
Instrumentation data of this kind at Green Peter was insufficient. The geological
information available is applicable to the uplift study.
4.4 Detroit Dam

4.4.1 Description
Detroit Dam is located on the North Santiam River 6 miles west and down-
stream of Detroit, Oregon, and about 75 miles southeast of Portland. The structure
was built in the early 1950s and consists of a concrete gravity dam and spillway
376 ft high above the controlled minimum tailwater, with a crest length of
1,579 ft. The highest monoliths are 15 through 21 in the center of the dam.
Monolith 19 is the highest.
Figure 4.20 is a plan of Detroit Dam. The dam is oriented northeast to south-
west with the forebay on the south side of the structure. An elevation of the dam,
looking downstream, is shown in Figure 4.21. Two sections through the dam are
presented in Figure 4.22. Section A-A is through the spillway, and Section B-B is
through the penstock. Both sections show locations of the grouting and drainage
gallery, where the instruments are accessible for monitoring. The dam is 450 ft
high from the top of excavated rock, forming a maximum crest elevation at
1,541 ft. The maximum width of the dam is 321 ft at the foundation-rock contact.
Foundation exploration began in June 1951. A total of 119 borings were

drilled and an exploration tunnel was driven. No boring logs or other critical
information was available for this dam. The foundation report, dated December
1952, described the geology but did not contain the boring logs that were refer-
enced in the report. A personal sketch provided by the Portland District contained
a section through the rock fracture system between monoliths 15 and 23 but did
not provide any other detailed information or dimensions. Portland District

Chapter 4
Figure 4.19. Locations of pressure measurement instruments, Green Peter Dam
81
Figure 4.20. Plan view of Detroit Dam

Figure 4.21. Elevation view of Detroit Dam

Figure 4.22. Sections through Detroit Dam
reported that the foundation rock consisted of andesite breccia, diorite, aplite,
andesite porphyry, and hydrothermally altered phases of these rocks. The degree
of fracturing rather than the hardness of the rock was the basis for determining its
adequacy for foundation purposes. Andesite breccia and diorite dominated, cut by
fissures and joints trending generally N 45o W (315o). The fissures were narrow
and tight.
Thirty-six uplift gauges were installed during construction; they have been
monitored continuously to the present. Piezometers are located in monoliths 9
through 24 and are concentrated mainly in monoliths 15 and 21 (Figure 4.23).
Of a total of 36 gauges, 23 piezometers are located coincidentally with the
maximum height of the dam. Dam periodic inspection brochures indicated that

Figure 4.23. Locations of piezometers, Detroit Dam

19 piezometers had no change in readings with time. An additional 11 piezo-
meters responded very irregularly and had no apparent relationship to fluctuation
of forebay levels. Seven piezometers provide useful data. Piezometers 15XXX,
16A, 16XX, 20A, and 20X showed a good response to forebay levels. Piezometer
17A was not very sensitive to fluctuations, but provided a good response. Piezo-
meter 15XXX exceeded the forebay at high pool. Piezometer 15A responded in a
way that could be tied to forebay fluctuation but had some unusual peaks.
Seven of the 36 gauges showed potential for consideration in the uplift uncer-
tainty study. Those piezometers, which provided consistently good readings, were
located in monolith 16 and possibly monolith 20. Because there were only two
piezometers in a cluster, there was no information on uplift through the cross
section of the dam, since the piezometers farther away from the forebay had no
readings. Instrumentation and geologic data were rated “insufficient.” Figure 4.24
shows instrumentation details.
4.5 Wolf Creek Dam

4.5.1 Description
Wolf Creek Dam is a combination earth-fill and concrete gravity dam. It is

located in Kentucky on the Cumberland River, 461 miles above the Ohio River.
Construction of Wolf Creek Dam began in August 1941, but it was discontinued
in August 1943 because of World War II. Construction of the dam and power-
house was resumed in September 1946, and the dam was completed in August
1951. The dam became operational in August 1952. It consists of a homogeneous
rolled-fill embankment 3,940 ft long and a concrete portion 1,796 ft long forming
an east wing of the project. The maximum pool is 158 ft (depth), and extreme
maximum pool is 195 ft. The estimated average pool is 152 ft. The normal design
pool is 135 ft. The spillway crest is at elevation 760 ft, and the minimum design
tailwater is at elevation 550 ft. The width of the dam at the concrete portion is
approximately 120 ft (varies with foundation rock elevation) at the highest mono-
lith. Figures 4.25 and Figure 4.26 are a plan and section of Wolf Creek Dam,
respectively.
The initial geological exploration predates World War II. Because of leakage
problems encountered after reservoir filling, more fundamental exploration and
mapping followed in the early 1960’s before remedial work was undertaken in
1968-1970. A wealth of borings resulting from extensive exploration was avail-
able for this dam. Borings adjacent or related to monolith 11 (which had potential
for the study) consisted of at least five borings: 2-RB at 90 ft upstream, 3-RB
coinciding with the monolith, 4-RB at 100 ft downstream, and 2- and 3-RV at

Figure 4.24. Details of instrumentation at Detroit Dam

Figure 4.25. Plan view of Wolf Creek Dam

Figure 4.26. Sections through Wolf Creek Dam
60 and 170 ft upstream, respectively. Locations and types of borings are presented
in Figure 4.27.
Foundation rock consists of argillaceous or shaly limestone and calcareous

shale. Initial geological exploration indicated that jointing controlled rock leakage.

Figure 4.27. Boring locations, Wolf Creek Dam

Numerous solution channels, caves, and small tunnels were uncovered during
excavation of the cutoff trench and were confirmed during additional exploration
in the 1960’s. The foundation of the concrete monoliths was grouted in three
stages: the first-stage grout holes reached to a depth of approximately 25 ft, the
second stage to a depth of 50 ft, and the third stage to a depth of approximately
75 ft into the foundation. The grout holes are on 5-ft centers along the dam’s
longitudinal axis. A system of drain holes on 10-ft centers parallel to the line of
the grouting holes was drilled 6 ft downstream of the line of grouting holes after
completion of grouting.
Geological information for Wolf Creek Dam was excellent, but instrumen-
tation was lacking. Most piezometers are located under the fill embankment
portion. An additional grout curtain was installed diagonally in the downstream
embankment adjacent to the concrete portion of the dam. Periodic inspection
reports documented leakage through the grout curtain. Remedial work was
undertaken on several occasions.
In 1972, six uplift cells were installed in Wolf Creek Dam’s grouting and
drainage gallery (Figure 4.28). Three cells were located upstream and three were
located downstream of the grout curtain. Only readings from Gauge M-11U fluc-
tuated with changes in the forebay until 1984, when the reading dropped to tail-
water level. Before the quick-release coupling method, coinciding with the sudden
drop of the gauge reading, was modified, no true readings were obtained at this or
any other gauge. The gauges probably were not connected long enough to reflect
true pressures in the instruments. In 1994, 14 additional piezometers were
installed to supplement existing cells. Like the original cells, they are angled 30°
upstream and downstream from the vertical axis of the dam. Only gauge M-11U
can be linked to variations in the forebay level. The remaining original gauges
follow tailwater trends. The newly installed additional gauges provide a flat plot
and do not follow fluctuations in either the forebay or the tailwater.
Figure 4.28. Uplift cells in Wolf Creek Dam

A single gauge record is considered a poor and insufficient data source for
uplift evaluation and made this dam a poor candidate for the uplift uncertainty
study. Geologic information for Wolf Creek Dam was rated “very good.”
Instrumentation data were rated “very sparse.”
4.6 Old Hickory Dam

4.6.1 Description
Old Hickory Dam is in Sumner and Davidson Counties at mile 216.2 on the
Cumberland River, approximately 10 miles northeast of Nashville, Tennessee.
Old Hickory Dam is a combination concrete gravity and rolled earth-fill structure.
It was built between 1952 and 1954. The earth embankment section is about
2,800 ft long. The dam is about 50 ft high, and its concrete portion is 30 ft wide at
the foundation base. Pool elevation fluctuates little (between 442 and 445 ft). The
highest pool elevation ever recorded was 450 ft. Figure 4.29 is a plan and one
embankment section of the dam. No cross section at the gravity part of the dam
was available.
The dam is within the regional upwarp of the Nashville Dome. A geological
investigation was undertaken in 1951 by drilling a total of 100 holes, and an
additional two holes in 1952. The deepest boring was 88 ft deep. There are
approximately 44 borings coinciding with the concrete portion of the dam, spill-
way, and lock (Figure 4.30). The deepest boring reaches about 40 ft into the rock.
At the dam site, drilling identified interbedded argillaceous and relatively pure
limestones. Some cavernous conditions were uncovered at the site of the spillway.
The upper 40 ft of rock was excavated for construction of the concrete portion of
the dam. Available boring logs did not provide information on rock fractures.
Only 12 borings coincided with locations of the uplift gauges. Constructing a
suitable geologic profile would be difficult.
Ten uplift gauges were installed in monoliths on the spillway during construc-
tion. Refer to Figure 4.31 for locations of the uplift instruments. All ten gauges
are read regularly. Periodic inspections reported that flow from most of the uplift
cells is very slow, requiring the measuring gauges to be left connected for a long
time. Of the ten uplift cells, readings for three cells did not change in 5 years, one
cell read pressure lower than the downstream water level, and the remaining cells
were constant. Piezometer O-13 reflects fluctuations of the tailwater. Piezometer
O-14 responds only to the tailwater. Piezometer O-7 shows some dependence on
tailwater and is reportedly strongly influenced by the upstream grout curtain.

Figure 4.29. Plan and section of Old Hickory Dam

Figure 4.30. Boring locations, Old Hickory Dam

Figure 4.31. Uplift instruments in Old Hickory Dam

Detailed drawings showing installation and exact locations of uplift cells were
not found in any available report or documentation. The uplift data history did not
indicate that any information consistent with the goals of the uplift study was
available. The dependence of gauge readings on the tailwater would obscure
rather than elucidate relationships of the forebay control mechanism on uplift
pressure effects induced by underseepage. Instrumentation was rated “insuffi-
cient.” The grouting curtain is located in such way that the uplift pressures and
direction of seepage through the substratum would be difficult to determine.
Geologic data for Old Hickory Lock and Dam were rated “insufficient” because
the essential parameters would be impossible to find or derive. Instrumentation
data for Old Hickory Lock and Dam were unsuitable for any serious modeling
effort.
4.7 J. Percy Priest Dam

4.7.1 Description
J. Percy Priest Dam was built between 1963 and 1968 on Stone River east of
Nashville Municipal Airport, Tennessee. Figure 4.32 is a plan and elevation
upstream of the dam with water impounded from the south. The dam is a combi-
nation earth-fill and concrete gravity dam. Only the central portion of the dam is a
concrete gravity structure. Two sections through the concrete gravity dam are
shown in Figure 4.33. The western part of the earth embankment is 1,340 ft long;
the eastern part is 622 ft long. The concrete gravity dam in the middle is 753 ft
long at the crest, 130 ft high, and about 80 ft wide at the foundation-rock inter-
face. The concrete portion of the dam consists of 15 monoliths varying in width
along the dam axis between 34 and 59 ft.
Geological exploration was performed extensively during 1929 and 1930.

Drilling continued in 1941, again from 1943 to 1944 and 1947 to 1948, with
additional drilling in 1962. Figure 4.34 shows locations of borings. A total of
109 borings, some inclined 45°, were emplaced during the three periods of
exploration. A detail of the geologic section at monoliths 4 and 5 is provided in
Figure 4.35. Relatively flat-lying, thin-bedded limestone with shale layers, some
very soft, was encountered. Drilling crews reported cavitation and water leakage
from neighboring borings under pressure during drilling. This required construc-
tion of a grout curtain, although no major caverns were uncovered during
construction.
Borings of importance to the uplift study are located near monoliths 5, 11, and
15. Exploration done in proximity of monolith 5 included borings 20-I and P-4.
Boring 6X-9 is downstream from the axis of the dam in monolith 5. Boring 6X-4
is about 60 ft from monolith 11. Monolith 15 has a larger number of borings:
calyx hole No. 1 and borings P-1, 6X-1, and 6X-2. Monolith 15 is adjacent to the

Figure 4.32. Plan and section of J. Percy Priest Dam

Figure 4.33. Sections through concrete portion of J. Percy Priest Dam

Figure 4.34. Locations of borings, J. Percy Priest Dam

100
Chapter 4
Figure 4.35. Geologic section, foundation of J. Percy Priest Dam

embankment and is less suitable for uplift modeling. Flow through the embank-
ment portion near monolith 15 would make modeling of flow very complex.
Figure 4.36 shows a geologic section along the axis of the dam, the Lower
Carters, Lebanon, and Ridley formations in relation to the locations of the
embankments and concrete gravity structure in the center.
During construction of the dam, six piezometers were initially installed in the
gallery: U-1, U-7, U-15, D-1, D-7, and D-15. The letters U and D designate the
locations of the piezometers within the monolith upstream or downstream of the
grout curtain. In 1983, an additional nine uplift cells were installed in a line per-
pendicular to the axis of the dam at monoliths 4, 5, and 11. Piezometers (uplift
cells) P-56 through P-59 are located in monoliths 4 and 5. P-60 through P-64 were
installed in monolith 11. All other piezometers in this dam, including four rock
piezometers, are located in the embankment near the trace of a fault. The embank-
ment piezometers have no relevance to the uplift study. All piezometers (uplift
cells) are inclined away from the grout curtain.
Initially, piezometer D-7 fluctuated closely with the tailwater reading. The
remaining piezometers did not fluctuate either with the forebay or the tailwater. In
November 1978, instruments in monolith 7 were damaged by a bomb explosion
and repaired. Readings taken since the repair do not indicate any changes at all,
plotting with a flat response. The piezometers installed in 1983 continue to show a
reasonable correlation with tailwater fluctuations, except for piezometers P-60 and
P-64, which show little or no response to either the forebay or the tailwater, and
plot a straight line. Figure 4.37 is a longitudinal section along the dam axis show-
ing the location of piezometers in the embankments and uplift gauges, shown as
U- and D- numbered devices, in the concrete portion. Figure 4.38 is a plan view
corresponding to the section in Figure 4.37. Locations of only four gauges are
shown in this figure, since some uplift gauges were added later and information
on those locations was not available.
For the purposes of the uplift study, uplift pressures related to tailwater
fluctuation have little value. Instrument readings depending on both the forebay
and tailwater would be too complex for interpretation. It would be difficult to
evaluate the effects of the various responses. The instrumentation was judged
unusable and insufficient. Geologic information also was insufficient.
4.8 Selection of a Case History

Documentation to support the selection process varied significantly among the
dams. In the process of consolidation of the Corps of Engineers District libraries,
many dam records related to dam construction have been transferred to the ERDC
technical library. Because of the disparity in age of the dams, records varied

102
Chapter 4
Figure 4.36. Geologic section along axis of J. Percy Priest Dam

Chapter 4
Figure 4.37. Locations of piezometers and uplift gauges, J. Percy Priest Dam
103
104
Chapter 4
Figure 4.38. Plan view of J. Percy Priest Dam showing locations of uplift gauges

greatly in quality. Historically, older dams have less supporting technical docu-
mentation than newer dams. Many project documents were not furnished to the
ERDC library, and under additional inquiry with the Districts, were not available
in the Districts’ libraries, nor in geology or foundation and materials branches of
the respective Districts. Full-scale copies of the borings (boring logs) were often
replaced in the foundation reports and other design memoranda by sketchy geo-
logical profiles. Fracture systems and sizes of fractures could not be evaluated, in
most instances, from available documentation.
Table 4.1 rates the evaluated dams for documentation essential to an uplift
study. They are listed in descending order of their potential evaluated by the pre-
ceding criteria. In the final stage of selection, the decision narrowed between
Libby Dam and Dworshak Dam. Both dams had suitable geologic information and
functioning uplift pressure monitoring instrumentation and data. Libby Dam was
selected for its greater magnitude of data recorded in the cluster of gauges within a
single monolith. Gauge readings at Libby Dam appeared more consistent and
linear for all gauges at all distances from the forebay to the toe. Repeated grouting
and extensive remedial work at Dworshak Dam relegated it to second choice.
Table 4.1
Summary of Uplift Data Assessment
Dam District State Geology Instrumentation
Libby Seattle Montana Good Good
Dworshak Walla Walla Idaho Good Sufficient
Detroit Portland Oregon Insufficient Insufficient
Green Peter Portland Oregon Not determined Possible
Wolf Creek Nashville Tennessee Very good Very sparse
Old Hickory Nashville Tennessee Insufficient Insufficient
J. Percy Priest Nashville Tennessee Insufficient Insufficient
4.9 Other Observations

Uplift pressure gauges installed in the foundations of the dams evaluated and
documented in this chapter were generally placed in the upper few feet of the
foundation rock, near the monolith-rock interface. Uplift pressure measurements
thus reflected pressures only in the upper part of the rock mass, and not neces-
sarily pressures in discontinuities at depth. Grout and drain holes were extended to
depths sufficient to intercept a number of potentially permeable discontinuities,
but uplift pressures that were monitored by gauges are not necessarily representa-
tive of pressures in those discontinuities. There is the potential for development of
unexpected uplift pressures in discontinuities deeper than the pressure gauges.
Sliding instability might then occur along the deep discontinuity rather than along
the concrete-rock interface.
Design engineers usually assume a linear, non-site-specific drained uplift

pressure distribution in the foundation (for example, Ebeling et al. 2000). Uplift
pressure gauges presumably monitor actual pressures developed in the foundation

following reservoir filling, but the preceding paragraph explained that critical
pressures in uninstrumented rock mass discontinuities may go undetected and
unheeded because of the shallow placement of uplift gauges. Assessments of
uplift pressure distribution and prediction in rock-founded concrete dams should
consider the possibility that existing pressure gauge systems may not monitor
critical pressures.

5 Hydraulic Properties of
Jointed Rock from
Pressure Tests at Libby
Dam
5.1 Background
Ebeling, Pace, and Morrison (1997) discussed procedures for predicting uplift
pressures beneath concrete dams founded on rock. Four procedures widely used to
predict uplift were (a) use of prescribed non-site-specific uplift distributions,
(b) computation from confined, one-dimensional (1-D) steady-state flow within a
rock joint, (c) computation of flow in a 1-D tapered rock joint, and (d) flow-net-
computed uplift pressures. They cited studies of existing dams (Stone and
Webster Engineering Corporation 1992) suggesting that foundation geology,
particularly the condition of rock joints, strongly influences the development and
distribution of uplift pressures beneath large gravity dams. Ebeling and Pace
(1996a) and Pace and Ebeling (1998) studied the effects of joint geometry and
aperture on the flow of water through the foundation and on the development of
uplift pressures. They used a 1-D steady-state laminar flow analysis and two-
dimensional finite element model to investigate the effects of joint aperture on
computed uplift pressures. These models incorporate estimates of joint aperture
and joint hydraulic conductivity in computing developed uplift pressures. In
assessing uncertainties in predicting uplift pressures, it is useful to attempt to
derive joint hydraulic properties from commonly available test data. This chapter
documents computations of hydraulic properties from borehole pressure test data
and borehole log information at Libby Dam obtained prior to dam construction.
The authors derived values of equivalent hydraulic conductivity1 of pressure-

tested zones of rock in the foundation of Libby Dam. The coefficient of equivalent
hydraulic conductivity is the hydraulic conductivity of a zone of jointed rock in
which flow is assumed to occur uniformly throughout the rock mass rather than
1
As explained in Chapter 2, the term hydraulic conductivity is used in this report in keeping with
recent usage. The terms permeability and coefficient of permeability are restricted to describing the
intrinsic property of a medium to transmit fluid independent of the fluid properties.
Chapter 5 Hydraulic Properties of Jointed Rock from Pressure Tests at Libby Dam 107
through individual joints1 (Bennett and Anderson 1982). Although flow actually
occurs through one or more joints in otherwise impermeable rock, treating the
rock as a uniform mass allows the use of Darcy’s law of diffuse flow in calcu-
lating an equivalent hydraulic conductivity for a given zone of rock in which
pressure tests have been conducted.
Foundation investigation borings placed in 1962 and 1963, prior to construc-

tion of Libby Dam, penetrated up to 120 ft of jointed argillite (a metamorphosed
claystone). About 60 ft of sand and gravel overlay the rock in most of the founda-
tion boreholes at the time of drilling. Seattle District pressure-tested many of the
boreholes by inflating packers to seal off approximately 10-ft zones of rock either
as the borehole progressed or after completion of a borehole. To conduct the tests,
District personnel pumped water into the test zones at pressures ranging from 30
to 100 psi and measured inflow in gallons per minute, converting to cubic feet per
minute. Pressure was measured at a surface gauge and inflow in a surface
flowmeter. For selected boreholes within monolith 23 of Libby Dam, the authors
applied Hvorslev’s equation for radial flow to a well through a permeable layer
between impervious strata to derive the equivalent hydraulic conductivity of each
pressure-tested zone.
From computed values of equivalent hydraulic conductivity and a count of the

number of joints in each test interval, equivalent joint aperture (also called equiva-
lent parallel plate or conducting aperture) and equivalent joint hydraulic conduc-
tivities were derived for each test zone using procedures suggested in the Rock
Testing Handbook (Geotechnical Laboratory 1993). Line graph plots of depth
versus derivative properties show the distribution of hydraulic properties within
the foundation of monolith 23.
5.2 Geological Conditions at Libby Dam

Libby Dam is located in northwest Montana on mile 221.9 of the Kootenai
River, 17 miles from its confluence with the Columbia River (Figure 5.1). It is a
concrete gravity dam 420 ft high and 3,055 ft long at its crest, impounding a
reservoir 90 miles in length that extends into Canada. Initial geological investi-
gations were conducted between 1951 and 1954; detailed foundation investiga-
tions were completed in 1966. The dam was completed in July 1973. The
maximum reservoir pool is elevation 2,459 ft. Libby Dam and its reservoir are in
the northern Rocky Mountains, a region characterized by rugged mountains and
linear valleys trending north to northwest. Maximum relief in the area is about
5,000 ft. The region is underlain by thick pre-Cambrian metasediments known as
the Belt Series, consisting of argillite, quartzite, metasandstone, and limestone.
As stated in Chapter 4, rock at the dam site, described in the foundation report
(U.S. Army Engineer District, Seattle, 1979), consists mainly of quartzite, meta-
sandstone, and siliceous argillite (a weakly metamorphosed mudstone). Rock
1
The term joint is used in this discussion to describe rock mass discontinuities in the foundation
of Libby Dam. It is understood that many of the “joints” are actually bedding plane separations.
108 Chapter 5 Hydraulic Properties of Jointed Rock from Pressure Tests at Libby Dam
Libby Dam
0 6
Scale, miles
Figure 5.1. Location of Libby Dam, Kootenai River, Montana
in the foundation of Libby Dam was described as hard, thin-bedded argillite with
sandy and calcareous zones. The foundation has a well-developed fracture system
consisting of several kinds and sets of joints, as described in the foundation report.
Figure 5.2 shows detailed mapping of discontinuities in the vicinity of mono-
lith 23. Bedding joints (congruent with the attitude of the bedded argillite) are
20 0 20 40
Scale in feet
Figure 5.2. Mapped discontinuities on foundation floor, monoliths 22, 23, and 24,
Libby Dam
common and strike 330°1 and dip 40° to 45° west. A prominent set of generally
east-west, high-angle shear joints dips 60° to 80° north or south. Where shear
joints intersect bedding joints, wedges were formed in the valley walls. Generally
north-south “relaxation joints” dip 50° to 80° east. Transverse “tension” joints
strike northeast and dip at moderate to high angles to the southeast. Other tension
joints strike parallel to bedding and dip at right angles to bedding and are prob-
ably related to folding. Many low-angle rebound joints (formed during unloading
of the rock mass) have random strikes. Some prominent bedding joints are filled
with gouge and are slickensided, evidence of movement, and are considered
faults. Several episodes of movement were noted along the faults. Intersection of
faults with east-west and north-south trending joints has broken the rock mass into
discrete blocks. Certain joints were open to a considerable depth, possibly caused
by unloading after glaciation.
Two impervious cutoff trenches were built for Libby Dam (U.S. Army Engi-
neer District, Seattle, 1979). The trenches did not reach the top of rock through
the full length of the cofferdam. To provide additional stability, the dam founda-
tion was excavated with a slightly lower elevation at the axis than at the down-
stream toe. Grout holes are inclined 25° upstream and 15° into the left abutment.
The primary grouting zone was 40 ft deep, the secondary zone 90 ft, and some
locations 160 ft deep.
Two drainage curtains assist the operation of the grout curtain: one 5 ft down-
stream and one 105 ft downstream of the grout curtain. The foundation report
stated that two lines of drains were considered necessary because of the 290-ft
base width of the dam. Drain holes varied from 60 to 105 ft in depth (U.S. Army
Engineer District, Seattle, 1979). An exploration adit was incorporated into the
drainage gallery.
Uplift pressure monitoring cells were installed under six monoliths: two
abutment monoliths, 14 and 41, and four valley monoliths, 18, 23, 29, and 34.
5.3 Computation of Equivalent Hydraulic

Conductivity
5.3.1 Principles and assumptions
Bennett and Anderson (1982) reviewed methods of determining flow proper-

ties in bored rock masses. They stated that an equivalent hydraulic conductivity
(they used the term equivalent coefficient of permeability) can be estimated using
data from constant head pressure tests if certain assumptions of groundwater flow
are satisfied:
a. The rock mass is homogeneous, isotropic, and saturated.

b. All flow is radial and axisymmetric about the borehole.
c. The borehole is vertical.
1
Structural attitudes are given in azimuth degrees, 0 to 360, clockwise from north.
d. Flow is steady state (at equilibrium, or constant head).
e. Flow is laminar.
f. A linear relationship exists between pressure and flow rate (i.e., Darcy’s
law is valid).
g. There is no leakage around the packer(s).
h. The change in pressure caused by acceleration of flow into the rock mass
is negligible.
When these conditions are met, the equivalent hydraulic conductivity Ke may
be calculated from results of constant-pressure tests using the equation derived by
Hvorslev (1951, his Figure 12, Case 9).1 Bennett and Anderson rearranged
Hvorslev’s equation to solve for Ke:
Ke = (Q/2πlH) ln (R/ro) (5.1)
where
Ke = equivalent hydraulic conductivity (units of L/T)
Q = volume flow rate at equilibrium (L3/T)
l = length of test section (L)
H = excess pressure head (L) at center of test section = Pt/γw + Hg, where Pt is
the pressure measured at the surface gauge and Hg is the head produced
by the height of water in the flow pipe (depth to the water table for a
submerged test section). Calculation of H neglects head losses between
the surface gauge and the test section and head due to flow velocity at
the gauge because these parameters were unknown or unrecorded
R = radius of influence of the pressure test (L)
ro = borehole radius (L)
The authors also reviewed the suggested method for pressure testing for the
determination of rock mass hydraulic conductivity, method RTH 381-80 in the
Rock Testing Handbook (Geotechnical Laboratory 1993). RTH 381-80 also
provided relationships for determining equivalent parallel plate aperture (e) and
joint hydraulic conductivity (Kj). Zeigler (1976) provides a thorough review of
procedures for determining rock mass hydraulic conductivity from pressure tests.
5.3.2 Libby Dam model
Borehole logs from the foundation report for Libby Dam (U.S. Army Engi-
neer District, Seattle, 1979) provided data for the pressure tests. Pressure-test
1
Hvorslev’s Case 9 is actually for a confined aquifer bounded above and below by impermeable
soil or rock. The Libby Dam foundation pressure tests were in zones confined only in the borehole
by inflatable packers, or by the bottom of the hole and a packer at the top of the zone. Flow within
the foundation rock could thus occur from or into rock above and below the sealed zone, unlike
flow in a permeable zone modeled by Hvorslev’s Case 9.
sections were nominally 10 ft long, the length of a typical drilling run. Some tests
were conducted as boring progressed, using a single inflatable packer in combina-
tion with the bottom of the hole to define each section. Other tests were conducted
after completion of a borehole, using a pair of packers to define a test zone, start-
ing at the bottom of the hole and progressing upwards. Pressure readings were
recorded on a gauge at the ground surface as #, presumably pounds per square
inch. Excess pressure heads (H, Equation 5.1) were therefore corrected for height
above the test section. Figure 5.3a illustrates the physical model describing the
case of the Libby Dam pressure tests. Figure 5.3b illustrates Hvorslev’s analytical
model applied to the pressure test data to derive rock mass hydraulic properties.
Most tests were run for 10 minutes with the flow (Q) measured for each 1-minute
interval. The value of Q used to calculate the hydraulic properties discussed in the
following paragraphs was the mean of the ten 1-minute measurements.
Zeigler (1976) reviewed Corps procedures for conducting pressure testing in

jointed rock masses. He stated that in most pressure tests the water injection
pressure is limited to a value that is not expected to increase the joint aperture (by
hydrofracturing). An increase in aperture would cause erroneously high flow rates,
resulting in higher and unrepresentative permeabilities than actually exist in the
rock mass. Common practice is to limit the water injection pressure to 1 psi/ft of
borehole depth above the water table and 0.5 psi/ft of borehole depth below the
water table. This criterion results in a maximum injection pressure less than the
effective overburden pressure if the overburden has a unit weight greater than
144 lbf/ft3 (overburden pressure 144 lbf/ft3 = 1 psi/ft dry, and 1 psi minus
0.43 psi/ft = 0.57 psi/ft submerged). Seattle District adjusted gauge pressures for
shallow test zones to maintain safe pressures in some of the pressure tests con-
ducted at Libby (roughly 0.5 to 1 psi per ft of depth: see column B in Table 5.1).
In others, however, no adjustment was made. For example, in borehole D-125,
gauge pressures were maintained at 100 psi in all 12 tested zones, including zones
as shallow as 66 ft depth.
Referring to Equation 5.1, Q is in cfm; l, H, Hg, R, and ro are in ft, Pt is in psi,

and Ke is calculated in ft/min. The pressure Pt is converted to ft of head by multi-
plying by 2.31 ft/psi.1. Hg is taken to be the depth from the ground surface to the
water table as recorded on the borehole log (all test sections were below the water
table). The borehole radius, ro, is (2.98 in./2)/12 = 0.124 ft (from the NX drill bit).
R, the radius of influence, is unknown; but because the quotient R/ro is a function
of the natural logarithm (ln), the value of ln (R/ro) does not vary significantly with
large changes in R (ln (R/ro) = 4.38, 6.69, and 8.99 for values of R = 10, 100, and
1,000, respectively). The effect on Ke of assuming an incorrect radius of influence
is not significant (Bennett and Anderson 1982; Zeigler 1976, p. 43). Because
there were many uncertainties associated with estimating equivalent hydraulic
conductivities for the Libby Dam foundation boreholes, including sparse data on
geometry and spacing of individual joints, the possibility of leakage through the
packers, distribution of flow within individual joints of a tested zone, and magni-
tude of flow velocities within individual joints, the importance of obtaining a
precise value for R was further reduced. Following Bennett and Anderson (1982),
1
Pt/γw = Pt/(62.4 lbf/ft3)/(144 in.2/ft2) = Pt * 2.31 ft/(lbf/in.2) = Pt * 2.31 ft/psi.
114
Chapter 5
Figure 5.3. Schematic of borehole pressure tests, Libby Dam, (a) physical model; (b) Hvorslev’s analytical model applied to pressure tests at Libby
Dam (after Hvorslev 1951)
Hydraulic Properties of Jointed Rock from Pressure Tests at Libby Dam

Table 5.1
Estimates of Borehole Interval and Joint Hydraulic Properties From Borehole Pressure Tests, Monolith 23, Libby Dam
Chapter 5
Borehole
Q, cfm Gauge, psi Hg, ft H, ft l, ft N Ke, ft/min Ke, cm/sec e, ft e, in. e, mm Kj, ft/min Interval, ft
Boring No. D-40, Libby Dam Monolith 23
0.56 85 8 204.35 12.1 17 1.58241E- 8.03864E-0 1.96702E- 0.0023604 0.0599548 0.5725931 69.0-81.1
Boring No. D-92, higher pressure
6.5 55 18 145.05 10.5 9 0.0029819 0.00151481 6.17171E- 0.0074060 0.1881137 5.6368768 65.0-75.5
6 70 18 179.7 10.5 10 0.0022217 0.00112867 5.40201E- 0.0064824 0.1646533 4.3185562 75.5-86.0

3.2 100 18 249 10.5 13 8.55169E- 4.34426E-0 3.60045E- 0.0043205 0.1097417 1.9184095 86.0-96.5
0.8 100 18 249 10.5 9 2.13792E- 1.08607E-0 2.56392E- 0.0030767 0.0781481 0.9728261 96.5-107.0
0 100 18 249 10.5 4 0 0 0 0 0 0 107.0-117.5

0 100 18 249 10.5 8 0 0 0 0 0 0 117.5-128.0
Boring D-92, lower pressure
4.8 30 18 87.3 10.5 9 0.0036587 0.00185862 6.60718E- 0.0079286 0.2013867 6.4604007 65.0-75.5
4.7 40 18 110.4 10.5 10 0.0028328 0.00143911 5.85775E- 0.0070292 0.1785441 5.0779566 75.5-86.0
2.3 50 18 133.5 10.5 13 0.0011464 5.82387E-0 3.96999E- 0.0047639 0.1210051 2.3324120 86.0-96.5
0.56 50 18 133.5 10.5 9 2.79131E- 1.41799E-0 2.80227E- 0.0033627 0.0854130 1.1621066 96.5-107.0

0.03 70 42 203.7 10.5 18 9.80013E- 4.97847E-0 7.28295E- 8.73953E- 0.0221984 0.0784949 64.5-75.0
0.02 80 42 226.8 10 21 6.16138E- 3.12998E-0 5.83101E- 6.99721E- 0.0177729 0.0503170 75.0-85.0
0 90 42 249.9 10.5 8 0 0 0 0 0 0 85.0-96.0
3.2 100 42 273 10.5 23 7.79990E- 3.96235E-0 2.88696E- 0.0034643 0.0879945 1.2334156 96.0-106.5
2.4 100 42 273 10.5 15 5.84992E- 2.97176E-0 3.02464E- 0.0036295 0.0921910 1.3538629 106.5-117.0
1.3 100 42 273 10.5 14 3.16871E- 1.60970E-0 2.52293E- 0.0030275 0.0768989 0.9419725 116.5-127.0
0.08 100 42 273 10.5 17 1.94997E- 9.90587E-0 9.33642E- 0.0011203 0.0284574 0.1289997 126.5-137.0
0.04 100 42 273 10.5 25 9.74987E- 4.95293E-0 6.51639E- 7.81967E- 0.0198619 0.0628407 136.5-147.0
0.34 100 42 273 10.5 12 8.28739E- 4.20999E-0 1.69850E- 0.0020382 0.0517703 0.4269330 146.5-157.0
(Sheet 1 of 4)
115
Table 5.1 (Continued)
116
Borehole
Boring No. D-93, lower pressure
2.2 50 42 157.5 10.5 23 9.29488E- 4.72180E-0 3.06074E- 0.0036728 0.0932912 1.3863705 96.0-106.5
1.1 50 42 157.5 10.5 15 4.64744E- 2.36090E-0 2.80131E- 0.0033615 0.0853839 1.1613159 106.5-117.0
0.6 50 42 157.5 10.5 14 2.53497E- 1.28776E-0 2.34208E- 0.0028104 0.0713866 0.8117673 116.5-127.0
0.14 50 42 157.5 10.5 12 5.91492E- 3.00478E-0 1.51790E- 0.0018214 0.0462655 0.3409682 146.5-157.0
Boring No. D184A, lower pressure
3.9 50 2 117.5 10 11 0.0023190 0.00117809 5.22275E- 0.0062672 0.1591893 4.0366941 51.0-61.0
4.5 50 2 117.5 10 20 0.0026758 0.00135934 4.48817E- 0.0053858 0.1367994 2.9810292 57.0-67.0

4.4 50 2 117.5 10 11 0.0026164 0.00132913 5.43703E- 0.0065244 0.1657207 4.3747296 67.0-77.0
2.2 50 2 117.5 10 13 0.0013082 6.64568E-0 4.08164E- 0.0048979 0.1244084 2.4654570 77.0-87.2
Chapter 5
D-184A, higher pressure
4.9 100 2 233 10 11 0.0014693 7.46440E-0 4.48577E- 0.0053829 0.1367262 2.9778406 51.0-61.0
5.6 100 2 233 10 20 0.0016792 8.53074E-0 3.84257E- 0.0046110 0.1171215 2.1850994 57.0-67.0
5.6 100 2 233 10 11 0.0016792 8.53074E-0 4.68994E- 0.0056279 0.1429494 3.2550881 67.0-77.0
3 100 2 233 10 13 8.99614E- 4.57004E-0 3.60270E- 0.0043232 0.1098103 1.9208102 77.0-87.2
Boring No. D-188
2 50 14.5 130 10.5 16 0.0010237 5.20058E-0 3.56733E- 0.0042807 0.1087322 1.8832771 63.0-73.5
0.1 90 14.5 222.4 10.5 5 2.99204E- 1.51995E-0 1.61927E- 0.0019431 0.0493553 0.3880311 73.5-84.0
0 100 14.5 245.5 10.5 12 0 0 0 0 0 0 84.0-94.5

0 100 14.5 245.5 10 1 0 0 0 0 0 0 89.0-99.5
Boring No. D-278
4.7 30 3.7 73 8.5 17 0.0052923 0.00268850 5.63374E- 0.0067604 0.1717162 4.6969995 50.6-59.1
3.3 75 3.7 176.95 13.6 15 9.58110E- 4.86720E-0 3.88638E- 0.0046636 0.1184568 2.2352076 56.3-69.9
0 75 3.7 176.95 12.8 10 0 0 0 0 0 0 66.3-79.1
0 75 3.7 176.95 11.3 23 0 0 0 0 0 0 76.8-88.1
(Sheet 2 of 4)

Table 5.1 (Continued)
Chapter 5
Borehole
Boring No. D-191
4.5 70 12 173.7 34.8 15 5.20145E- 2.64234E-0 4.33639E- 0.0052036 0.1321731 2.7828134 50.0-84.8
6.6 70 25 186.7 10.5 5 0.0023523 0.00119499 6.93688E- 0.0083242 0.2114361 7.1212496 62.5-73.0
6.7 80 25 209.8 10.5 4 0.0021250 0.00107953 7.22366E- 0.0086683 0.2201773 7.7222292 72.7-83.2
7 90 25 232.9 10.5 3 0.002 0.001016 7.79154E- 0.0093498 0.2374862 8.9841018 82.7-93.2
2.33 35 25 105.85 10.5 5 0.0014647 7.44098E-0 5.92361E- 0.0071083 0.1805515 5.1927789 62.5-73.0
5.3 40 25 117.4 10.5 4 0.0030040 0.00152606 8.10719E- 0.0097286 0.2471071 9.7267627 72.7-83.2
5.5 45 25 128.95 10.5 3 0.0028381 0.00144180 8.75577E- 0.0105069 0.2668759 11.345314 82.7-93.2
Boring D-125, higher pressure (12 tests)
8.7 100 5 236 10 11 0.0025757 0.00130846 5.40870E- 0.0064904 0.1648570 4.3292518 66.0-76.0
8 100 5 236 10 16 0.0023684 0.00120318 4.64202E- 0.0055704 0.1414888 3.1889066 63.0-73.0

0.23 100 5 236 10 6 6.80937E- 3.45916E-0 1.97201E- 0.0023664 0.0601068 0.5755011 76.0-86.0
0.27 100 5 236 10 11 7.99360E- 4.06075E-0 1.69971E- 0.0020396 0.0518070 0.4275391 86.0-96.0

0.4 100 5 236 10 2 1.18424E- 6.01593E-0 3.42027E- 0.0041043 0.1042498 1.7312048 96.0-106.0
0.27 100 5 236 10 16 7.99360E- 4.06075E-0 1.50014E- 0.0018001 0.0457242 0.3330358 106.0-116.0
0.27 100 5 236 10 10 7.99360E- 4.06075E-0 1.75457E- 0.0021054 0.0534794 0.4555866 116.0-126.0
0.27 100 5 236 10 13 7.99360E- 4.06075E-0 1.60765E- 0.0019291 0.0490010 0.3824800 126.0-136.0
0.53 100 5 236 10 21 1.56911E- 7.97110E-0 1.71555E- 0.0020586 0.0522898 0.4355448 136.0-146.0
1.2 100 5 236 10 17 3.55271E- 1.80478E-0 2.41710E- 0.0029005 0.0736732 0.8646028 146.0-156.0
1.41 100 5 236 10 12 4.17444E- 2.12061E-0 2.86459E- 0.0034375 0.0873128 1.2143775 156.0-166.0
1.73 100 5 236 10 20 5.12183E- 2.60189E-0 2.58656E- 0.0031038 0.0788383 0.9900851 166.0-176.0
(Sheet 3 of 4)
117
118
Table 5.1 (Concluded)
Borehole
Boring D-125, lower pressure (11 tests)
6 50 5 120.5 10 11 0.0034790 0.00176733 5.97877E- 0.0071745 0.1822328 5.2899408 66.0-76.0
0 50 5 120.5 10 6 0 0 0 0 0 0 76.0-86.0
0 50 5 120.5 10 11 0 0 0 0 0 0 86.0-96.0
0.13 50 5 120.5 10 2 7.53784E- 3.82922E-0 2.94214E- 0.0035305 0.0896764 1.2810143 96.0-106.0
0 50 5 120.5 10 16 0 0 0 0 0 0 106.0-116.0
0 50 5 120.5 10 10 0 0 0 0 0 0 116.0-126.0
0 50 5 120.5 10 13 0 0 0 0 0 0 126.0-136.0
0.27 50 5 120.5 10 21 1.56555E- 7.95301E-0 1.71425E- 0.0020570 0.0522502 0.4348853 136.0-146.0
Chapter 5
0.8 50 5 120.5 10 17 4.63867E- 2.35645E-0 2.64184E- 0.0031702 0.0805231 1.0328545 146.0-156.0
0.67 50 5 120.5 10 12 3.88489E- 1.97352E-0 2.79677E- 0.0033561 0.0852455 1.1575528 156.0-166.0
0.87 50 5 120.5 10 20 5.04456E- 2.56264E-0 2.57349E- 0.0030881 0.0784398 0.9801020 166.0-176.0
Boring D-126, higher pressure
5.25 60 15.5 154.1 10 55 0.0023803 0.00120923 3.08096E- 0.0036971 0.0939076 1.4047502 72.7-83.5
5.72 60 15.5 154.1 10 55 0.0025934 0.00131749 3.17028E- 0.0038043 0.0966302 1.4873857 72.7-83.6
0.16 85 15.5 211.85 10.5 29 5.02566E- 2.55303E-0 1.07132E- 0.0012855 0.0326538 0.1698499 72.7-83.7
0.25 100 15.5 246.5 10.5 20 6.74877E- 3.42838E-0 1.33778E- 0.0016053 0.0407756 0.2648491 72.7-83.8
0.03 100 15.5 246.5 10.5 12 8.09852E- 4.11405E-0 7.82340E- 9.38808E- 0.0238457 0.0905771 72.7-83.9
4.16 30 15.5 84.8 10 55 0.0034275 0.00174121 3.47909E- 0.0041749 0.1060427 1.7912642 72.7-83.12
4.01 30 15.5 84.8 10 55 0.0033039 0.00167842 3.43676E- 0.0041241 0.1047525 1.7479420 72.7-83.13
0.07 50 15.5 131 10.5 29 3.55573E- 1.80631E-0 9.54620E- 0.0011455 0.0290968 0.1348618 72.7-83.14
0.08 50 15.5 131 10.5 20 4.06369E- 2.06435E-0 1.12967E- 0.0013556 0.0344322 0.1888551 72.7-83.15
(Sheet 4 of 4)

suggesting a value for R of between l and l/2, a value of 10 ft for R was selected,
and the resulting value for ln (R/ro) of 4.390059 was used in the computation of
Ke.
Zeigler (1976) evaluated the “zone of influence” around a borehole during

pressure testing. He calculated the drop in pressure head at distance from the
borehole during radial flow and showed that a 50 percent head loss occurs within
a radial distance of about 3 to 4 ft from the borehole. Citing theoretical and lab-
oratory studies, he also showed that large changes in the radius of influence, R, do
not greatly affect the rate of head drop. During a pressure test there is a severe
drop in pressure near the borehole. The area over which a pressure test is effective
is unknown but may well be within only a few feet of the test section. Conse-
quently, only those joints intersecting the borehole will materially influence the
test results (Zeigler 1976, p 50).
5.3.3 Pressure tests versus pumping tests
Pressure tests in rock differ from pumping tests in that pressure tests are of
short duration and affect a much smaller volume of the rock mass than do pump-
ing tests (Headquarters, U.S. Army Corps of Engineers, 1993, pp 3-10 and 3-11).
A pumping test injects water into or withdraws water from a well at a constant or
variable rate for a considerable period of time and measures the drawdown of the
piezometric surface in observation wells within the aquifer. Pumping tests involve
large volumes of the rock mass and tend to average the effects of discontinuities
within the rock mass. A pressure test, which pumps water into a well under con-
stant pressure and measures the resulting flow rate, is of short duration and affects
only a small volume of the rock mass because frictional losses in the immediate
vicinity of the test section are commonly large. Pressure tests therefore provide
more accurate information on the effects of discontinuities near the borehole.
5.3.4 Pressure testing at Libby Dam
Figure 5.4 shows the locations of foundation boreholes in the vicinity of

monolith 23, Libby Dam. Summary logs for boreholes are presented as Fig-
ures 5.5 through 5.12. The author reviewed archived files at the Seattle District
office to obtain original or facsimile copies of the full-sized borehole logs, pres-
sure test data, and other field documents. Pressure-test data provided in the
“Remarks” column of each summary log were augmented by the field logs. In
many cases, additional pressure-test information, including tests at lower or higher
pressures, was available in the field logs.
Pressure tests were generally 10 minutes in duration. Some boreholes were

tested at two pressures that differed by a factor of approximately two. Usually, but
not always, tests at the higher pressures were conducted first. Water inflow from
the boreholes into the rock mass during the pressure tests (in the monolith 23 area)
ranged from 0 to 8.7 cfm. Gauge pressures ranged from 30 to 100 psi. Pressures
were not always the same for different test intervals within a borehole.
D109
568400
D108
D184A
D125 Line of section,
Figure 5.21
568200 D183A
D75
Monolith 23 D126
Northing, State Plane, FT
D92
568000
D188
D40
D278
D93
567800 D191
D94
567600
0 100 200
Scale, ft
567400 D113
588800 589000 589200 589400

Easting, State Plane, FT
Figure 5.4. Locations of boreholes, monolith 23 area, Libby Dam. Star denotes
borehole was pressure-tested
Figure 5.5. Summary log of boring D-92, monolith 23 area, Libby Dam
5.3.5 Results for Libby Dam model
Data required for Equation 5.1 and for Equations 2.13 and 2.5 and calcula-
tions of joint and rock mass hydraulic properties for eight boreholes were com-
piled in an ExcelTM spreadsheet, a printout of which is presented as Table 5.1. Ke
is tabulated in ft/min and cm/sec, e in ft, in., and mm, and Kj in ft/min. Graphs of
Ke, N (number of joints intersecting the test section), e, and Kj versus depth were
then plotted. The graphs are presented as Figures 5.13 through 5.20. Where
pressure tests at two pressure ranges were conducted for a borehole, the graphs
show Ke versus depth for both pressures. Computations of e are presented in the
next section.
5.4 Computation of Parallel Plate Aperture and

Joint Hydraulic Conductivity
Studies of the effects of uplift on sliding stability at Bluestone Dam, West
Virginia (Fuller, Mossbarger, Scott, and May 2000), suggested that examination
of mechanical apertures was not sufficient to predict flow into and through a frac-
ture. Investigators relied instead on packer (pressure) tests and laboratory tests of
conducting aperture for estimates of conducting aperture, e. For the Libby Dam
exercise, hydraulic conductivities of equivalent individual joints were estimated
using the procedure suggested in RTH 381-80 (Geotechnical Laboratory 1993)
and in Zeigler (1976) whereby pressure test data are first used to compute a
parallel plate aperture, e. The test section is assumed to be intersected by a group
of parallel and identical joints. Each joint is assumed to be an equivalent parallel
(smooth-walled) plate. Flow to the borehole is assumed to be radial and to occur

only within the joints. The rock mass between the joints is assumed impermeable.
Joint hydraulic conductivity, Kj, is then estimated using e and physical properties
of the fluid. Equations 2.13 and 2.5 (rearranged) are applicable (see Chapter 2 for
discussion):
Kj = (e2γw)/12 µw (2.5 rearranged)
The number of joints, N, for a test interval was determined by counting the joints
in the graphic borehole log. Applying Equations 2.13 and 2.5, e and Kj were
estimated for each pressure-tested zone in the ten boreholes analyzed for mono-
lith 23. Using Equation 2.13, an estimate of conducting aperture e was computed.
The value of µw in Equation 2.13 is 3.479 E -5 lbf min/ft2 × 0.0101 = 3.5138

E -7 lbf min/ft2.1 The value of 12µw/γw is 12(3.5138 E -7)/62.4 lb/ft3 = 6.7573
E -8 min ft.
With Equation 2.5, the equivalent individual joint hydraulic conductivity Kj

was calculated by squaring the value of e derived by Equation 2.13 and multi-
plying it by the reciprocal of 12µw/γw. The results of applying Equations 2.13
and 2.5 are presented in Table 5.1, columns I, J, and K (e in ft, in., and mm,
respectively) and L (Kj in ft/min). See Figures 5.13 through 5.20 for plots of e
versus depth for eight of the ten boreholes.
1
Dynamic viscosity of water at 20oC = 0.0101 poise. 1 poise = 1 dyne sec/cm2. 1 dyne = 2.247
E -6 lbf. So 1 poise = 2.247 E -6 lbf, X (1/60)min/[1/(30.482)]ft2 = 3.479 E -5 lbf min/ft2. Dynamic
viscosity of water = 3.5138 E -7 lbf min/ft2.
Figure 5.10. Summary log of boring D-184A, monolith 23 area, Libby Dam
Chapter 5
129
Figure 5.13. Equivalent hydraulic conductivity, number of joints per test interval, equivalent joint aperture, and equivalent joint hydraulic
conductivity, borehole D-92, monolith 23, Libby Dam. Data points are at centers of test intervals
130
Chapter 5

Chapter 5
131
132
Chapter 5

Chapter 5
133
134
Chapter 5

conductivity, borehole D-184A, monolith 23, Libby Dam. Data points are at centers of test intervals
Chapter 5
135
136
Chapter 5

5.5 Discussion of Results
5.5.1 Equivalent hydraulic conductivity, Ke
Cursory examination of the plots of Ke versus depth (Figures 5.13-5.20),

suggests a general decrease in equivalent hydraulic conductivity of the rock mass
with increasing depth. Next, depths were converted to elevations for comparison
of Ke with increasing depth (decreasing elevation) in all boreholes. Figure 5.21, a
composite plot of all values of Ke versus elevation (all boreholes), confirms that
the higher equivalent hydraulic conductivities occur near the top of the rock mass.
The top of rock in the vicinity of the boreholes ranged from about 2,072 to
2,086 ft MSL. Data for Figure 5.21 are for tests in each of nine boreholes in
which pressure tests were conducted (data are for only the highest pressure range
in boreholes in which two tests were conducted). In a rock mass in which ground-
water flow is through discrete discontinuities, higher hydraulic conductivities are
expected in the upper portion of the rock mass where weathering and stress relief
of joints have created greater apertures. Equivalent hydraulic conductivities
decrease rapidly below about elevation 2,045 ft MSL. Some tight joints (near-zero
Ke) occur throughout the rock mass. The cluster of four data points showing a
slight increase in Ke with increasing depth from elevations 1,986 to 1,956 ft are
for a single borehole (borehole D-125). Plots of conducting aperture, e, versus
elevation and versus depth are presented later in this section.
The central plot of Figures 5.13 through 5.20 shows the number of joints
(joint frequency) logged in each borehole at depths representing the centers of the
pressure test intervals. The distribution of joints with depth appears to be random.
That is, there is no apparent correlation of joint frequency with depth below top of
rock. In most boreholes, there is little correlation between joint frequency and
equivalent hydraulic conductivity Ke. The rightmost plot of Figures 5.13 through
5.20 shows the distribution of equivalent joint apertures computed for each pres-
sure test interval. There is a reasonably good correlation between e and Ke in the
plots, but little or no correlation between e and the number of joints. For example,
Figure 5.13, borehole D-92, shows a wide variation in joint counts between 80 ft
and the bottom of the borehole, but a relatively constant decrease in aperture e and
conductivity Ke with increasing depth. Figure 5.14, borehole D-93, shows a
dramatic increase in joint aperture accompanying a sharp rise in Ke between 90-
and 120-ft depth, but with a lower joint count in the same interval. Recall that
equivalent aperture e is derived from Ke using the cubic law and the joint count, N
(see Equations 5.1 and 2.11). The implication is that joint aperture, not joint
frequency, is the dominant factor in determining zones of high flow in jointed
rock masses. Of course, joints must also be persistent enough or sufficiently
interconnected that a flow path is sustained to the tailwater. That is, joints must
not “dead-end” short of the outlet.
Borehole data for the vicinity of monolith 23 were used to construct a cross
section through the monolith foundation (Figure 5.22). Borehole data are pro-
jected into the section (note that boreholes do not necessarily lie within mono-
lith 23, as shown in Figure 5.4). The cross section shows selected logged features
in those portions of the boreholes below top of rock (as defined by exploratory
2080
Elevation of center of test interval, ft MSL 2060
2040
2020
2000
1980
Each point represents a

single pressure test in one
1960 test interval in one of a total
of nine boreholes
1940
0 0.002 0.004 0.006

Equivalent hydraulic conductivity, Ke, ft/min
Figure 5.21. Variation in equivalent hydraulic conductivity with elevation in nine

boreholes near monolith 23. From foundation pressure tests, Libby
Dam
boreholes). The bold line near the center of the section is the profile of the dam
(monolith 23). Vertical enhancement of the section is 5 to 1. Graphs of Ke beside
each borehole log show the distribution of computed hydraulic conductivities in
pressure-tested boreholes within the foundation. The dashed line delimits the base
of a zone of relatively high hydraulic conductivity as defined by the pressure test
data and subsequent calculations. The base of the conductive zone is essentially
the elevation at which the computed conductivities dropped below about
10-3 ft/min. The zone coincides generally with regions of “open” joints or
“shattered” rock in boreholes. The zone of higher conductivity is thicker upstream
and in the upstream portion of monolith 23 than in the downstream portion. Exca-
vation of the foundation removed up to about 18 ft of weathered and more perme-
able rock prior to pouring of the monolith (see the bold line of the monolith in
Chapter 5
D-92

Figure 5.22. Geologic cross section through foundation of monolith 23, Libby Dam (location shown in Figure 5.4). Borehole data are projected into
the section. Logs show locations of discontinuities in boreholes. Graphs show equivalent hydraulic conductivity (Ke) from pressure-
139
test data at approximately 10-ft intervals. Section is perpendicular to dam axis, downstream to right
Figure 5.22). The cross section indicates that permeable rock extends another
15 ft or so deeper than the base of the excavation in the vicinity of the heel of the
dam and upstream.
An anomaly became apparent when values of Ke were computed for boreholes

in which pressure tests were conducted at two distinct pressures. Equivalent
hydraulic conductivities from tests at high pressures were consistently lower than
conductivities computed from tests at the lower pressures. The difference was
greatest at higher values of Ke. Figure 5.13, the plot of Ke versus depth for bore-
hole D-92, illustrates the phenomenon. Computed values of Ke for the 55- to
100-psi tests were consistently lower than values computed for the 30- to 50-psi
tests, and the difference was consistently greater with increasing Ke. The differ-
ence is most prominent for values of Ke greater than 10-3 ft/min. Figure 5.23
shows that a relationship exists between the difference in Ke and the magnitude of
Ke. The plot has two implications. First, the range in the difference in computed
Ke at two pressures is linearly related to the magnitude of the computed Ke.
Second, there is more scatter in the range in computed values of Ke between the
test pairs at higher magnitudes of Ke. This plot may signify that there is simply
more error inherent in computations of equivalent hydraulic conductivity at higher
flow rates and in larger aperture joints than in less permeable joints.
What is more difficult to explain is the tendency for the higher pressures to
consistently result in lower computed Ke than the lower pressures. Referring to
Table 5.1 (Borehole D-92, higher pressure and lower pressure), higher pressures
(gauge psi, column B) produced, as expected, mean higher flow rates (Q, cfm,
column A) at equivalent borehole intervals (column M) in the pair of tests (about
33 percent higher than in the low-pressure tests). Gauge pressures in the high-
pressure tests were about double that of the low-pressure tests (mean 91 percent
higher). H, the excess head (column D, Table 5.1) consists of gauge pressure, in
ft, plus gravity head Hg. H was a mean 77 percent higher in the high-pressure tests
than in the low-pressure tests. Computed Ke, however, was lower at the higher
pressures than at the lower pressures by a mean 26 percent. Equation 5.1 shows
that Ke is directly proportional to flow rate Q and inversely proportional to excess
pressure head H. If Q and H change proportionally from one test to another, Ke
changes proportionally. In the monolith 23 pressure-test pairs, Q (the numerator in
Equation 5.1) changed much less than H (the denominator) from high-pressure to
low-pressure tests. The result was that high-pressure tests produced lower values
of computed Ke than low-pressure tests, which intuitively should not be the case if
the relationship between hydraulic conductivity, Q, and H is linear (i.e., Ke com-
puted from high-pressure tests should be the same as that computed from low-
pressure tests). In other words, high-pressure tests produced a lower inflow (Q)
than low-pressure tests (in open joints).
These results may signify that turbulent, or nonlinear, flow is occurring in the
joints at higher pressures. Equation 5.1 implies that Ke is linearly proportional to
flow rate Q. Zeigler (1976) and Headquarters, U.S. Army Corps of Engineers
(2001) state that flow rates that are not proportional to pressure in a test zone may
indicate turbulent flow. It is possible that turbulent flow at the higher pressure
impedes the flow of water through the test interval and results in a lower Q and
lower computed Ke. Nonlinear flow at higher test pressures is a source of potential
Chapter 5
Figure 5.23. Relationship of equivalent hydraulic conductivity (Ke) of pressure-tested zones to difference (range) in Ke for test pairs, Libby Dam.
Data are for pairs of tests at two pressures, P1 and P2, where P1 = approx. 2*P2 (see key for explanation)
141
error in computations of Ke and conducting aperture e and leads to uncertainty in
predicting corresponding uplift pressures in numerical models.
5.5.2 Conducting aperture, e
Although the graphs in Figures 5.13 through 5.20 show a clear tendency
toward decreasing e with depth, a simple statistical report was made of all aper-
tures treated independently of depth (see Table 5.2). Equivalent parallel plate
apertures e computed with Equation 2.11 from the higher pressure test data ranged
from 0 to 0.2375 mm. Maximum computed e was 0.2375 mm (238 microns
(µm)). Mean computed e was 0.0817 mm (82 µm). The standard deviation, an
indication of the spread of the e values about the mean, was 0.0769. Perhaps a
better measure of the spread of values is the coefficient of variation (the standard
deviation divided by the mean), which expresses the standard deviation as a
percentage of what is being measured. The coefficient of variation was 0.7673, a
high number that indicates that the standard deviation is almost as high as the
mean and implying a wide variation in computed values of e.
The computed parallel plate aperture, e, is not the true joint aperture but the
conducting aperture. As explained in Chapter 2, the conducting aperture e is the
distance between two smooth parallel plates that would allow the same flow as a
mechanical aperture (E) with rough walls. Conducting aperture is always less than
or equal to mechanical aperture (Ebeling, Wahl, and Pace 1997). Computed
values of e at Libby Dam are in the same range as apertures computed for dams by
other investigators (Barton, Bandis, and Bakhtar 1985; Snow 1965).
Table 5.2 The relationship of conducting

Statistical Information on aperture, e, and depth is shown in
Computed Values of Equivalent Figures 5.24 through 5.26. Fig-
Aperture, e ure 5.24 shows the variation in e
with elevation in the foundation of
Number of values 49
Libby Dam. A relationship similar
Minimum e, mm 0 to that between Ke and elevation is
Maximum e, mm 0.2375 apparent (higher apertures at
Range in e, mm 0.2375 higher elevations). Snow (1968)
Mean e, mm 0.0817
investigated rock fracture open-
ings (conducting apertures com-
Median e, mm 0.0769
puted from pressure tests) at
Standard deviation, mm 0.0627 several dam sites. He plotted
Coefficient of variation 0.7673 apertures against the depth at
which they occurred below the top
of rock or below the base of the
overburden. Snow presumably believed that joint and bedding planes opened
following erosion of the rock (denudation and subsequent stress release) and that
apertures should therefore be adjusted for depth below top of rock rather than for
depth from ground surface or for elevation. Apertures computed for Libby Dam
were adjusted for depth below top of rock for comparison with Snow’s results for
other dams. Figure 5.25 is a plot of computed apertures beneath monolith 23
versus depth below top of rock. The scale in Figure 5.25 is identical to that of
2080
Elevation of center of test interval ,ft MSL
2040
2000
1960
test interval in one of a total
of nine boreholes
100 microns 200 microns
1920
0 0.1 0.2 0.3

Joint conducting aperture, e, mm
Figure 5.24. Joint conducting aperture, e, versus elevation, Libby Dam

monolith 23
Figure 5.24 (the plot of e versus elevation). Figure 5.26 is a plot of Snow’s (1968)
data. Data for Libby Dam and for Snow’s dams are strikingly similar. Snow’s
conclusion that joint “…openings [aperture] decrease with depth…” and that
“…the marked decrease of openings with depth is most responsible for decreases
in permeability…” is consistent with the Libby Dam data. Note that the computed
conducting apertures for Snow’s discontinuities and for the Libby Dam joints are
also very similar, ranging from near zero to 200 or 300 µm. A marked decrease in
aperture occurs at about 40 to 50 ft in depth in both Snow’s and the Libby Dam
data. Apertures generally are less than 100 µm below that depth.
0
40
Depth of test interval below top of rock,ft
80
120
test interval in one of a total
of nine boreholes
100 µ 200 µ
160
0 0.1 0.2 0.3

Joint conducting aperture, e, mm
Figure 5.25. Joint conducting aperture, e, versus depth below top of rock, Libby
Dam monolith 23
In the case of the Libby Dam data, there is little difference in the shape of the
plots of e versus elevation and e versus depth below top of rock (Figures 5.24 and
5.25, respectively). Snow’s presumed concern about the need to reference aper-
ture to top of rock rather than ground surface is less a factor for the Libby Dam
monolith 23 area because the total relief on the original rock surface is only 14 ft
and on the original ground surface only 12 ft.
Using the relationship of Barton, Bandis, and Bakhtar (1985), Figure 2.8, an
expected equivalent mechanical aperture, E, for a conducting aperture, e, of
100 µm and greater would be 2 to 3 times the value of e (i.e., E/e = 2 to 4),
assuming a roughness (JRC) of 15 (typical value cited in Ebeling, Wahl, and Pace
1997). Maximum computed mechanical, or actual, apertures for monolith 23
Figure 5.26. Average rock fracture openings (in microns) versus depth below
overburden (top of rock) at dam sites (after Snow 1968, Figure 7)
joints would thus be in the range of 500 to 700 µm (0.5 to 0.7 mm). Following
work by Barton (1982, pp. 65-68), E could also be calculated from Equations 2.9
or 2.10 if measurements of JRC (from tilt tests, for example) were available {E =
[e/(JRC)2.5]0.5}. Examples of apertures describing open joints were presented in
Chapter 2. A joint was generally described as open if its mechanical aperture was
equal to or greater than 250 to 500 µm (see Table 2.1). Snow (1968) used a
conducting aperture as low as 35 mm to describe an open joint. Under those prece-
dents, most of the joints in the upper 25 to 50 ft of foundation rock in monolith 23
are open joints.
Barton (1982) suggested the use of borehole pressure tests to estimate con-
ducting and real apertures to predict joint deformation from future stress pertur-
bation of foundations and excavations. He advised using low injection pressures
to avoid reducing levels of existing effective stress and subsequent enlarging of
joint apertures (hydrofracturing). Barton stated that calculations of e from pressure
tests, and E from estimates of JRC, could be the starting point for estimations of
mechanical and hydraulic response of joints to further stress. His considerations
emphasize the need to define the uncertainties inherent in deriving discontinuity
characteristics from field tests that are not carefully controlled. Pressure testing in
some of the monolith 23 boreholes at Libby Dam was not limited to the suggested
0.57 psi/ft (discussed earlier in this chapter), and slight widening of existing joints
may have occurred in the upper portions of the rock mass.
6 Analysis of Uplift
Pressures at Libby Dam
Libby Dam is a concrete gravity dam. It consists of 47 monoliths, each

between 59 and 66 ft wide as measured along the longitudinal dam axis. The
uplift pressure gauges were installed in six monoliths: 14, 18, 23, 29, 34, and 41.
Monoliths 23 and 29 were considered for the dam uplift study because they are
centrally located and the hydrostatic pressure on the reservoir base is at its maxi-
mum in these monoliths. The basis for the selection of the monolith 23 uplift
pressure gauges was the quality and higher number of useful readings in its
available 20-year monitoring history.
The uplift pressure gauges are located at the two lowest galleries, mutually
parallel at elevations varying between 2,177.5 ft and 2,197.75 ft. These are the
“drainage and grouting gallery” located 5 ft downstream from the dam axis, and
“downstream drainage gallery,” 105 ft downstream from the dam axis
(Figure 6.1).
The grouting curtain incorporated into the dam foundation is composed of

three zones: a tertiary zone 40 ft into rock, a secondary zone 90 ft into rock, and a
primary zone 160 ft into rock. The grout holes top at the floor of the upstream
drainage and grout gallery. A section and plan of the grout holes and drains and
locations of uplift pressure instruments (gauges) are shown in Figure 6.1. The
grout holes are, as in all other valley monoliths, inclined 25° upstream and 15°
toward the left abutment as measured from vertical. The grout holes are on 5-ft
centers along the longitudinal axis of the dam.
The dam has a system of drains organized in two rows. The drains at the
upstream grout and drainage gallery are spaced evenly 10 ft apart parallel to the
dam axis in a row 10 ft downstream from the dam axis. All drains in monolith 23
at the upstream grout and drainage gallery are 105 ft deep. Drains outlet at the
level of the grout and drainage gallery. A profile view of this gallery is provided
in Figure 6.2.
A second row of drain holes is 95 ft downstream from the front row of the
drains. These drains are 60 ft deep. They outlet at the downstream drainage
gallery. All drain holes in both the upstream grout and drainage gallery and
downstream drainage gallery were drilled on 10-ft centers through 3.5-in. I.D.
galvanized pipes, each 5 ft long and embedded in concrete prior to construction.
Chapter 6 Analysis of Uplift Pressures at Libby Dam 147

Axis of the dam
Upstream gallery drains

{inclined 2%%d towards forebay}
Drains of downstream Forebay
{drainage gallery}
EL. 2080.0
EL. 2077.5 ft
piezometers (gages)
Toe of
Dam Grout
Section @ rock contact {10'}

25°
0 Scale in Feet 100
To Monolith 24
{80'} {60'} {40'} {20'} {20'} {20'} {10'} {10'}
L5 L4 L3 L2 L1
{20'}
Monolith 23
C8 C7 C6 C5 C4 C3 C2 C1
{63'}
{20'}
R5 R4 R3 R2 R1
To Monolith 22
{5'}
Legend: piezometers {10'} {10'}
drains
grouting holes Varies
PLAN AT EL. 2077.5

Instrument and Drain Locations at Monolith 23
Figure 6.1
Figure 6.1. Instrument and drain locations, monolith 23, Libby Dam
All drain holes in both galleries are 3 in. in diameter and inclined on an angle 15°
off the vertical toward the left abutment and 2° off the vertical toward the forebay.
A profile view of the downstream drainage gallery is provided as Figure 6.3.
A total of 18 pressure cells (gauges) were installed in monolith 23. The pres-
sure cells are organized in groups of three gauges in five rows spaced in equal 20-
ft intervals, with three additional pressure cells centered in unequal distances
downstream from the forebay group. The first row of gauges is 10 ft upstream of
the dam axis, the second row is 10 ft downstream of the dam axis, the third row is
30 ft downstream of the dam axis, the fourth row is 50 ft downstream of the dam
axis, and the fifth row is 70 ft downstream of the dam axis.
The gauges in this main group are designated L for gauges 20 ft left of the
center of the monolith, C for the center monolith gauges, and R for gauges 20 ft to
the right of the centerline of the monolith, facing the forebay. Three additional
pressure gauges are in the center of the monolith: gauge C6 is 110 ft downstream
from the axis, and gauges C7 and C8 are 170 ft and 250 ft, respectively, down-
stream of the axis of the dam. Refer to the plan in Figure 6.1 for respective
148 Chapter 6 Analysis of Uplift Pressures at Libby Dam

{63'} {63'}
17+71
16+45
17+08
22 23 24
Drains Grout
@ 10' center
EL. 2080.0 @ 5' center
EL. 2077.5 +
_ 3'
{5'} {Concrete}
{Rock}
{15°}
Drain
Grout
0 10 20 30 40 50 60 70 80 90 100
Scale in Feet
Upstream Grout Drainage Gallery

Figure 6.2. Upstream grout and drainage gallery, monolith 23, Libby Dam, looking
upstream
locations of the pressure gauges. A detail of the gauge installation at the rock/
monolith interface is shown in Figure 6.4.
6.1 Calculation of Uplift Pressures from Gauge

Readings
The Geology Section of the Seattle District furnished a complete set of data
files containing uplift pressure gauge readings, recorded forebay elevations in feet,
volume discharge, recorded tailwater elevations in feet, and ambient forebay

16+45
17+08
22 23 24
{5'} {5 @ 10'} {8'} EL. 2078.52

EL. 2080.0
2 3 4 5 6 1 2 3 4 5 6 1 2 3
Top of
Rock
{15°} {Drains}
20 10 0 10 20 30 40 50
Scale in Feet
Figure 6.3. Downstream drainage gallery, monolith 23, Libby Dam, looking
upstream
temperature. The hydraulic readings of forebay elevation were recorded on a daily

basis for all dates since November 1979. The uplift pressure gauge readings were
taken continuously on a monthly basis from October 1979 until the present. For
peak forebay elevations, the uplift pressure (gauge) readings were taken twice a
month.
The readings of the gauges represent total head in feet, which includes pres-
sure head and elevation head. The readings do not include velocity head, which is
considered negligible for this study. No conversion factor was indicated for the
data files obtained from Seattle District.
6.2 Variation of Uplift Pressure with Time

Two periodic inspection reports were available for this report: Periodic
Inspection Report No. 10 (1987) and Periodic Inspection Report No. 13 (1993).

Figure 6.4. Detail of uplift cell and installation, Libby Dam
The Seattle District provided a complete plot of gauges for the interval 1981-1999
for each monolith equipped with pressure gauges at Libby Dam. Plots for mono-
lith 23 gauges are included as Figures 6.5 through 6.12. Gauge C1 was replaced
in May 1995 and in December 1997. Gauges C2, C3, and C4 were replaced in
May 1993. All gauges were bled in March 1993, September 1995, April 1996,
July 1996, March 1997, and April 1998. A C1 bleeder valve was installed in May
1995. Gauges C5, C6, C7, and C8 were replaced in May 1993, and the gauges
were bled in March 1995, September 1995, April 1996, July 1996, March 1997,
August 1997, and April 1998. Gauges R1, R3, R4, and R5 were replaced and
gauge R2 bled in May 1993, and gauge R5 was replaced in September 1995.
Gauge R1 bleeder valve was installed in May 1995 and R3 gauge was replaced in
July 1996 and March 1997. All R gauges were bled in March and September
1995, in April 1996, in March and August 1997, and finally in April 1998.

152
Chapter 6
Figure 6.5. Uplift pressure at monolith 23, Libby Dam, L1-5, 1981-1990
Analysis of Uplift Pressures at Libby Dam

Chapter 6
Figure 6.6. Uplift pressure at monolith 23, Libby Dam, L1-5, 1991-1999
153
154
Chapter 6
Figure 6.7. Uplift pressure at monolith 23, Libby Dam, C1-4, 1981-1990
Chapter 6
155
156
Chapter 6

Chapter 6
157
158
Chapter 6
Figure 6.11. Uplift pressure at monolith 23, Libby Dam, R1-5, 1981-1990

Chapter 6
Figure 6.12. Uplift pressure at monolith 23, Libby Dam, R1-5, 1991-1999
159
Gauges L1 through L5 were replaced in May 1993 and bled in March and
September of 1995, in July 1996, in March and August 1997, and in April 1998.
Gauge L1 was replaced twice, in March 1995 and in May 1995, when a bleeder
was also installed.
Figures 6.5 through 6.12 are the time plots of the gauge readings in mono-
lith 23 and forebay readings for 1981 through July 1999. During the period 1981
through 1989, of all L gauges, only L1 gauge readings matched changes in fore-
bay elevations. Gauges L2, L3, L4, and L5 did not reflect the forebay fluctuation
in any consistent manner. Beginning in year 1989, gauges L3 and L4 more nearly
followed the trends of the forebay fluctuation. After mid-1990, gauges L1 and L4
closely matched the forebay elevation changes. Gauge L2 followed forebay
changes correctly after year 1991, except for the time between 1992 and 1994.
Gauge L3 followed the trends of the forebay until the early part of 1995,
becoming completely irregular afterwards. Gauge R1 showed close similarity in
readings to the forebay fluctuation for the whole period of 1981-1999. Gauge R2
provided a zero reading, displaying a value for the elevation of the tip of the
gauge. Gauges R3, R4, and R5 matched fluctuation of the forebay only in the
rising pool phase of the forebay cycle, showing a rapid increase in readings during
the first 9 years of the records. However, this portion of the yearly cycle was
followed by an immediate rapid decrease of gauge readings although the forebay
still maintained a high pool elevation. The readings became more regular after
1989 until 1994 when readings of gauge R5 became constant for 3 years. Since
the summer of 1996, gauges R3 through R5 showed a good match between their
readings and the forebay elevations.
Of all C gauges, only gauge C1 records matched forebay elevations during the
period 1981-1984. Gauge C2 recorded zero pressures for the whole period 1981-
1999. After 1985, gauge readings of C3 and C4 followed forebay fluctuations
very closely. Gauge C6 gave zero reading for the entire available period of
recording. Gauges C5, C7, and C8 behaved irregularly until the summer of 1985.
Gauge C8 remained irregular with some vague response to forebay fluctuation
until 1999. Gauges C5 and C7 gave good readings between 1989 and the spring
of 1993 when only C7 continued with reasonably good response to forebay
fluctuation.
For further analysis, a set of plots for gauges L1, C1, and R1 were fitted for
each yearly cycle between 1981 and 1999. The graphs consisted of recorded
forebay elevation, in feet, plotted on the horizontal axis and respective gauge
readings, in feet, plotted on the vertical axis. Every gauge reading of L1, C1, and
R1 was plotted for a complete 1-year cycle, separately for each gauge. Each cycle
represented an increasing and decreasing forebay water elevation. The resulting
graph showed a very flat hysteretic curve.
Two representative years, based on time records in Figure 6.5, were selected
to investigate changes in cyclic behavior in the selected time interval. The first
complete yearly cycle occurred in 1990 (Figure 6.13). For comparison, a similar
plot for year 1999 is included as Figure 6.14. An increase in forebay hydrostatic
pressure was reflected in a proportionate increase in respective gauge-monitored

2480.00
FOREBAY VS C1
FOREBAY VS L1
2440.00
FOREBAY VS R1
Gage Reading, Ft.
2400.00
2360.000
2320.00
2280.00
2280.00 2320.00 2360.00 2400.00 2440.00 2480.00
Forebay Reading, Ft.
Figure 6.13. Uplift pressure at monolith 23, Libby Dam, for calendar year 1990
pressures. A decrease of the forebay hydrostatic pressure as a result of pool

emptying was reflected in proportionately lower monitored gauge pressures on all
three gauges, L1, C1, and R1. The rising portions of the curves plotted very close
to the descending portion of the curve.
The second observation made from the plots in Figures 6.13 and 6.14 was
with respect to the slopes of the obtained curves. The slope did not increase or
decrease with the increased forebay pool level, but was constant. The third obser-
vation made in these two figures was that curves for respective plots of gauges L1,
C1, and R1 were almost identical to plots for years 1990 and 1999.
The last observation made from Figures 6.13 and 6.14 was the mutual loca-
tion of the plots for L1, C1, and R1. The plots for each of these gauges were not
identical but were slightly offset and parallel.

2480.00
2440.00
FOREBAY VS L1
Gage Reading. Ft.
2400.00
FOREBAY C1
2360.00
2320.00
FOREBAY VS R1
2280.00
2280.00 2320.00 2360.00 2400.00 2440.00 2480.00
Forebay Reading, Ft.
Figure 6.14. Plot of dependence of L1, C1, and R1 (in ft) on forebay elevation (ft), monolith 23, Libby
Dam, for calendar year 1999
6.3 Variation in Uplift Pressure Within the

Foundation
Figures 6.15 and 6.16 plot the distributions of the recorded uplift pressures
across monolith 23. Year 1990 data are plotted in Figure 6.15 and year 1999 data
are plotted for comparison in Figure 6.16. Only four selected readings were
included in the plots: for maximum forebay pool, for minimum forebay pool, and
for two intermediate stages of the forebay pool for both years. Gauges L1, C1, and
R1, in the first row of the forebay, responded very closely to fluctuations of the
forebay pool elevation in both years. Readings in the first row of gauges exceeded
design uplift pressures in both years, 1990 and 1999.

Elevation in Feet
Date (Elevation)
2475
7-23 (2458.32)
2450
2425
6-5 (2407.49)
2400
2375 2-14 (2369.73)
2350
4-12 (2329.37)
2325
2300
Design Hydraulic 2275

Uplift
2250
2225
2200
Tailwater 2175 Forebay

2-14 (2124.2)
2150
7-23 (2120.29)
2125
4-12 (2119.73) 2100
2075
6-5 (2117.94)
C8 C6 L2 L1
C7 L5 L4 L3 Grout Curtain
Drains
Figure 6.15. Recorded uplift pressures in 1990 versus design pressures, monolith 23, Libby Dam
Of the gauges in the second row, L2, C2, and R2 , only gauge L2 recorded
nonzero pressure readings. Gauge L2 readings are explained in the next section.
Gauge C6 showed a zero reading for all dates.
The design uplift pressure (shown in Figures 6.15 and 6.16 as a solid line
above the dotted lines) much exceeded the recorded gauge pressure. Uplift pres-
sure at gauge L3 was lower than gauge L4, even though gauge L4 is farther from
the forebay. Gauges C7 and C8 gave readings above the tailwater level, and the
gauge C8 reading was higher than the reading for gauge C7. There is no signifi-
cant difference in the uplift pressure profiles of Figure 6.15 and 6.16. There was
no noticeable difference between data collected in 1990 and 1999 at the various
gauges.
Table 6.1 sums the observation at individual gauges for 1990 and 1999.
The average pool elevation of the forebay was 5 ft higher in 1990 than in
1999, explaining why all readings of gauges in 1999 were lower than those taken
in 1990. As shown in Table 6.1, relative to forebay pool elevation, the gauge

Elevation in Feet
Date (Elevation)
2475
9-3 (2455.12)
2450 7-7 (2436.48)
2425
1-6 (2398.84)
2400
2375
2350
3-4 (2329.04)
2325
2300
Design Hydraulic 2275

Uplift
2250
2225
2200
Tailwater 2175 Forebay
1-6 (2124.7) 2150

3-4 (2122.6)
2125
9-3 (2121.7) 2100

7-7 (2122.2)
2075
C6 L2 L1
C8 C7 L5 L4 L3
Grout Curtain
Drains
Figure 6.16. Recorded uplift pressures in 1999 versus design pressures, monolith 23, Libby Dam
readings are very stable in a span of 10 years. The most controlling effect is the
location of gauges with respect to the grout curtain and to drains in the drainage
and grout gallery. Gauge L1, like the other gauges in the first row (C1 and R1), is
located between the forebay and grout curtain (Figures 6.15 and 6.16). The effect
of the drains at this location is minimal. Gauge L2 coincides to a lesser extent
with drains in the drainage and grout gallery, but is affected much less than
gauges C2 and R2 in the same row and at the same distance from the dam axis.
Gauge L2 occasionally reads a positive (nonzero) value, but in the same row
installed gauges R2 and C2 always yield zero readings. Gauge L2 may read non-
zero data at peak to average pool forebay elevations. Readings in gauge L4 exceed
those in gauges L3 and L5, since the latter are closer to drains on both sides than
gauge L4 (Figures 6.15 and 6.16). Gauge C6 yields no meaningful data. Readings
in gauge C7 farther from the forebay exceed readings from gauge L5, since gauge
L5 is closer to the second row of the drains (at the drainage gallery). The readings
on gauge C8 exceed the tailwater elevations.

Table 6.1
Pressure Observations in 1990 and 1999
1990 1990 1999 1999
Range Average Range Average
Reading Reading Percent Reading Reading Percent
Site ft ft Forebay ft ft Forebay
Forebay 2,329-2,459 2,394 100 2,321-2,458 2,389 100
L1 2,328-2,458 2,393 99 2,320-2,458 2,387 99

L2 2,095-2,118 2,106 88 2,096-2,112 2,104 88
L3 2,109-2,130 2,120 89 2,109-2,128 2,118 89

L4 2,105-2,142 2,124 89 2,107-2,144 2,126 89
L5 2,099-2,120 2,110 88 2,098-2,116 2,107 88

C7 0 0 0 0 0 0
C7 2,112-2,128 2,120 89 2,110-2,127 2,118 89
C8 2,132-2,146 2,139 89 2,111-2,137 2,124 89

Tailwater 2,118-2,125 2,122 89 2,118-2,125 2,122 89
6.4 Significance of Uplift Variations

A comparison of Figures 6.13 and 6.14, representing the relationship between
forebay pool elevations and gauge readings (in ft), for L1, C1, and R1, suggests
that no significant changes in readings of these gauges took place between 1990
and 1999. The almost identical slope of the hysteretic curves in both plots sug-
gests that increased forebay elevation causing higher hydrostatic pressure in front
of the dam does not cause accelerated increase of the uplift pressure. The pressure
recorded at higher forebay readings is proportional to the pressure at lower fore-
bay readings. There is no downward shift in the curve location in the graph
between 1990 and 1999. This supports the assumption that the gauges are fully
functioning.
Gauge L1 (Figures 6.13 and 6.14, purple) recorded higher readings than
gauge C1 (green), which recorded slightly higher readings than gauge R1 (red).
The difference is not substantial and can be explained by the particular location of
each of these three gauges with respect to the location of drains. The spacing of
the drains at the foundation/rock interface is not exactly the same as the spacing of
the gauges, resulting in a different collective effect of the drains on each of the
front row gauges.
There are three curves in each of Figures 6.13 and 6.14. They are similarly
shaped, and the loading side and unloading side of the hysteresis (forebay rising
and forebay lowering) of the gauge responses are very similar, or almost parallel.
This fact suggests that the gauge reading is, for all practical purposes, independent
of the direction of pressure increment. The underground flow is not accelerated
out of proportion when the forebay pool peaks.
Figures 6.15 and 6.16 provide a comparison of the design uplift pressure and
recorded uplift pressures in a cross section of the dam in years 1990 and 1999.

The figures show that expected uplift pressure was significantly overestimated at
the time of design on all gauges L2 through C8. Because of its position near the
line of drain holes, gauge L2 might be expected to reflect just the elevation head
(i.e., to reflect full drainage). Figures 6.17 and 6.18 help explain why gauge L2
shows a pressure higher than elevation head. The right side of Figure 6.17 is a
cross section through the first row of drain holes, which are shown as black lines
angled from the vertical. Gauge L2 is in a vertical hole the base of which is offset
from the angled drain hole. At the position of the gauge, the arching of the piezo-
metric surface between drain holes (projected along the red line in Figure 6.17)
causes the gauge to read a positive pressure in addition to the elevation head. A
three-dimensional diagram in Figure 6.18 provides another perspective on the
shape of the piezometric surface around the drains. The idealized surface in
Figure 6.18 assumes that the foundation is relatively homogeneous and that the
spacing and apertures of rock discontinuities are uniform throughout the monolith
foundation.
The grouting curtain and drains in both galleries were perhaps much more
effective than expected. The uplift pressures recorded on gauges between drains
correctly peak at gauge L4, located in the middle between drains. This further
supports the explanation given in Figures 6.17 and 6.18 in the case of gauge L2.
Recorded uplift pressures on gauge C8 indicate that the uplift pressure at the
tailwater exceeds the hydrostatic pressure of the tailwater for higher stages of the
tailwater.
6.5 Summary
Monolith 23 of Libby Dam has a total of 18 pressure gauges that are in good
condition and that are supplying and have supplied useful readings over the dam’s
20-year monitoring history. The Seattle District was able to furnish the authors a
complete set of data containing uplift pressure gauge readings, recorded forebay
elevations, volume discharge, and recorded tailwater elevations. These data per-
mitted a thorough evaluation of the response of the foundation to reservoir level
fluctuations and to changes in response with time.
Responses of installed uplift pressure gauges in monolith 23 of Libby Dam

were linear for the monitoring period 1990 to 1999. Linear response is an indica-
tion that gauge readings are independent of the direction of pressure change.
Linear response also implies that foundation discontinuities have a sufficiently
high aperture that reservoir loading and unloading do not significantly affect flow
through them.

Chapter 6
Figure 6.17. Section through drains, monolith 23, Libby Dam
167
Figure 6.18. Effect of drains on piezometric surface (idealized)

7 Conclusions and
Recommendations
7.1 Conclusions
Uplift is a major force affecting the stability of rock-founded concrete dams.
The flow regime within a rock foundation controls uplift pressures. Both the
geological interpretation and the analytical procedures used to calculate flow
within the foundation produce uncertainty in the calculation and prediction of
uplift pressures. Uncertainty in modeling uplift pressure manifests itself in three
areas: geologic uncertainty, material uncertainty, and spatial uncertainty.
7.1.1 Geologic uncertainty
Geologic uncertainty arises in mapping and describing the stratigraphy,

geologic structure, and degree of weathering that characterize a foundation.
Questions to be asked when gathering geologic information at the dam site
include the following:
a. Does the stratigraphy cause rock properties to vary within the

foundation?
b. Are the stratigraphic sequence of rock units and their corresponding
hydraulic properties adequately described?
c. How does weathering affect the hydraulic conductivity throughout the
foundation? Careful planning and execution of site investigation address
these issues. The engineering geologist should have a preconceived idea,
a conceptual model, of what the site geology is before beginning site
investigation. The conceptual model is then continuously updated as new
information becomes available. In this way, investigators know where
data are lacking and can take steps to fill in where needed to describe
adequately the foundation rock mass.
d. Have all joint sets been adequately described and sampled? Has bias in
drill hole alignment failed to sample a particular orientation? Vertical
boreholes, for example, fail to intercept most joints of a vertical joint set.
Alignment of exploration borings should be adjusted to ensure that all
sets are sampled.
Chapter 7 Conclusions and Recommendations 169

7.1.2 Material uncertainty
Material uncertainty pertains primarily to estimates of the hydraulic con-

ductivity of the rock mass. These are mostly field-test uncertainties from data
scatter in measurements caused by natural variations in material properties and in
problems with test precision. Thoroughness of site investigation is another source
of material uncertainty, for example:
a. Is the engineering geologist sure that the rock itself does not account for
significant flow in addition to flow through rock discontinuities? That is,
is rock primary porosity (intergranular porosity) contributing to flow?
This is one area where laboratory permeability tests on core samples
would be useful in quantifying rock hydraulic conductivity as
distinguished from rock mass conductivity through discontinuities.
b. Has site investigation isolated or detected the major flow conduits or flow
zones in the rock mass? Have major joint sets or areas of enlarged
openings such as karst solution passages been identified and charac-
terized? In other words, is the field investigation thorough? Preparing a
conceptual model of site geology with emphasis on locations of ground-
water flow paths and quantities of flow helps assure that field investiga-
tions are adequately designed. Structural and stratigraphic cross sections
showing hydraulic characteristics of mapped geologic units should be
prepared and continuously modified as field investigations proceed.
Field measurements of groundwater properties must address several issues:
a. Are assumptions of flow theory met? Is there laminar, radial, and

symmetric flow to the borehole in borehole aquifer tests? Is the rock mass
sufficiently homogenous and isotropic to warrant the use of Darcy’s law
in computing aquifer properties? Do boundary conditions, such as the
presence of impermeable barriers to flow, exist that would alter the shape
of the drawdown curve in a pumping test? These issues should be
addressed to determine their impact on numeric estimates of hydraulic
conductivity, joint aperture, and other uplift modeling parameters.
b. Are field procedures of aquifer testing consistent, and are measurements
of pressure and flow accurate? Personnel conducting field tests in bore-
holes should exercise quality control in test procedures to ensure that tests
are conducted in the same way in every borehole and that measures taken
to maintain accuracy of test results are followed consistently. Personnel
should test flow gauges and pressure monitoring devices periodically as
part of quality control.
c. Do pressure or aquifer test packers leak, and are piezometers or pressure
gauges working properly? Packer pressures should be monitored for leak-
age during pressure testing. Piezometers should be tested periodically to
eliminate these kinds of data errors as a source of uncertainty.
d. In pressure testing, has hydrofracturing been prevented? Water injected
at pressures exceeding overburden pressure may widen or open existing
discontinuities and produce artificially high flows or turbulent flow
170 Chapter 7 Conclusions and Recommendations

during pressure testing. Subsequently calculated values of hydraulic
discontinuity and joint aperture will be in error. Personnel should limit
injection pressures to about 0.5 psi per foot of depth for test zones below
the water table and 1 psi per foot of depth for tests above the water table.
e. How valid is the reduction of zone hydraulic conductivity (Ke) to specific
joint hydraulic conductivity (Kj) and derived values of conducting
aperture, e, in pressure tests? How reliable are the pressure test data;
i.e., is there enough redundancy in the tests to measure statistical vari-
ation as a source of uncertainty? These issues might best be addressed in
the design and execution of field pressure tests. Tests should be designed
with sufficient redundancy to measure the variation in measured quanti-
ties in a given test interval or for a specific joint being tested. Joint
hydraulic conductivities and conducting apertures derived from pressure
tests on packer-isolated rock mass zones can be compared with field or
laboratory tests on individual joints to determine the degree of
discrepancy in calculated and measured values.
7.1.3 Spatial uncertainty
Spatial uncertainty is represented in the ways properties vary throughout the

foundation. Measurements of material properties and geometries of geologic
features are made at point or line sources (boreholes), in narrow drifts, or along
rock exposure faces. These observations are then extrapolated beyond the sampl-
ing point or interpolated between two or more sampling points and applied to the
entire foundation rock mass, with accompanying uncertainty. The number of mea-
suring and sampling points should be adjusted for site complexity to accommo-
date the amount of variation expected as a result of stratigraphy, geologic structure
(folding, jointing, and faulting), and the effects of differential weathering.
The characteristics of rock mass discontinuities can be spatially variable.

Issues include the following:
a. Do joints persist; i.e., is there a high or low degree of “connectivity” or

persistence of joints across the dam base, particularly from the reservoir
to the tailwater?
b. Do joint conditions change with distance? These questions might be
addressed with careful and complete core logging of boreholes for dis-
continuities and zones of discontinuities. Careful core logging and
complementary observations of the borehole itself using borehole camera
techniques may permit correlation of specific discontinuities or jointed
intervals between boreholes. Well-planned rock mass pressure tests of
specific joints or jointed intervals between adjacent boreholes, for
example, in existing drainage galleries of the dam might provide
information on which joints or zones persist through the foundation.
There is also a temporal uncertainty inherent in dam foundation properties.

Joint conditions may vary with time because of changes in joint aperture caused
by stresses of loading and unloading by the reservoir, and by filling in or washing
out of materials between the joint walls. Temporal uncertainty is difficult to

address because it is usually not practical to bore and sample the foundation after
dam construction. Conducting pressure testing in existing drainage galleries or
other open borings may be one way to determine if joint hydraulic properties have
changed. Observation of joint conditions within the boreholes using borehole
imaging devices is another potential method to monitor changes in rock mass
discontinuities.
7.2 Recommendations
7.2.1 Geologic data for uplift modeling
The authors recommend a four-step approach to addressing issues of geologic

uncertainty in modeling of uplift pressures in rock foundations. The emphasis is
on ascertaining hydraulic properties and the condition of rock mass discontinuities
as they control distribution of flow and uplift pressures within the foundation.
Figure 7.1 illustrates the four steps in the assessment process, explained in the
following paragraphs.
4 Computation Steps
Locate & size rock discontinuities
Field test for joint properties
Construct a flow model
Compute uplift pressures
Figure 7.1. Recommended four-step approach to modeling of uplift pressures in

rock foundations
The first step is to locate and describe rock discontinuities. The joint system,
including bedding plane discontinuities, should be mapped and numbers and
orientations of joint sets determined. Tight joints, i.e., joints with mechanical
apertures less than about 250 µm, may cause nonlinear pressure conditions in the
foundation. Tight joints are more likely to deform sufficiently under reservoir
loading to profoundly affect joint hydraulic conductivity and to produce a

nonlinear distribution of pressure from heel to toe because of closing of the joints
at the toe. Numerical modeling must address the issue of whether linear or non-
linear conditions exist. Predominant joint apertures should be determined through
borehole measurements of joint apertures, for example, using borehole-imaging
devices, or through measurements of JRC and conducting aperture.
Joint condition (filling, degree of weathering, roughness) should be described

using similar borehole measurement methods. A conceptual model, with a repre-
sentative geologic cross section through the foundation, should be developed to
reduce site geology to a workable degree of complexity for numerical modeling.
For example, critical zones of discontinuities or specific discontinuities, changes
in lithology or stratigraphy that might affect flow or distribution of pressure, and
areas of weathered rock should be included in the section to define the conceptual
model.
This report documented an exercise in estimating joint hydraulic properties

(Kj, e, and E) from pre-existing data from pressure tests conducted routinely
during the foundation exploration phase of the construction of Libby Dam. While
results from this exercise were promising (reasonable values that agreed with
published data for other dams), they could be greatly augmented with carefully
planned and executed follow-up field tests in the foundation. The next step in the
four-step approach addresses this issue.
The second step in addressing uplift modeling uncertainties is to conduct field

pressure tests and borehole observation at a rock-founded damsite. The objective
of this exercise on a Corps of Engineers dam would be to perform additional field
testing to develop in sufficient detail an assessment of the rock discontinuity
regime and the parameters needed to perform a site-specific uplift evaluation. In
addition, the uncertainty in the parameters would be assessed. Tests should be
designed to determine the hydraulic conductivity, persistence, and conducting
apertures of specific discontinuities or zones of discontinuities. Testing of this
kind has been conducted at Bluestone Dam (FMSM 2000). Ideally, hydraulic
straddle packers would be used to isolate a specific joint in two or more existing
boreholes in the foundation. Libby Dam presents a unique opportunity to investi-
gate joint persistence and flow properties because its foundation is equipped with
two rows of drainage holes, as discussed in Chapter 6 of this report. Individual
drain holes are 10 ft apart, parallel to the dam axis, in each row, and the two rows
are 100 ft apart perpendicular to the axis. The goal of such field tests would be to
develop and refine techniques for extracting flow properties from existing dams
that require uplift prediction.
In tests within a single drain row, pressure tests could be conducted between
two or more adjacent boreholes to determine the hydraulic conductivity and con-
ducting aperture of specific joints or group of joints within the boreholes. Because
the boreholes are only 10 ft apart, there is a high likelihood that the same joint
could be identified in each hole. A pair of straddle packers in a central hole would
isolate a joint and serve as the injection boring. Pressures in the adjacent hole or
holes would be monitored during the pressure test to determine the pressure drop
and define the boundary condition of the test. Flow rate and injection pressure
would be measured in the central hole. Equations similar to those discussed in

Chapters 2 and 5 of this report would be applied (adjusted for specific test
boundary conditions) to extract values of Ke, Kj, e, and E. These numbers can be
applied directly to numeric modeling of uplift pressures. By working with a small
test area or volume, redundant tests can be used to isolate problems with test pro-
cedures and to define the natural variation inherent in test results. Redundancy is
necessary to obtain enough test data for statistical analysis of variations in test
results that are not caused by material (rock mass, groundwater) variations.
In tests between two rows of drain holes (at Libby Dam), separated by a
distance of 100 ft in a direction parallel to flow, the desired result would be a
determination of joint persistence. That is, how far does a single joint or set of
joints maintain flow through the dam foundation. Pressure tests between two
boreholes in different drain rows might reveal where hydraulic connectivity exists
between joints or groups of joints isolated by packers. If it is possible to isolate
multiple zones with more than one set of straddle packers in a borehole, pressure
or flow could be monitored in the downstream line of drains while a selected zone
is pressurized in the upstream line. There would be a better chance of detecting
flow, and persistence, between lines with multiple monitoring zones than simply
trying to correlate a specific joint or jointed interval over the 100-ft distance
separating the two rows of drains.
Supplementary data that would complement data from the pressure tests
would come from careful borehole imaging logging of the existing open drain
holes in the foundation. Selected drain holes would be logged using a high-
resolution borehole camera or other imaging device that permits measuring of
joint aperture on the borehole wall and observation of joint condition. Correlation
between boreholes of features in the borehole image log would be critical to
selecting test intervals for straddle packer pressure tests.
An interesting procedure performed in the Bluestone Dam data test program

consisted of laboratory hydraulic conductivity tests on core samples containing a
single joint. A small-diameter hole was drilled in the center of the core sample
from the top surface to just below the joint to be tested. The core sample was
placed in a permeameter device equipped with a loading platen for varying the
normal stress applied to the joint. The annulus surrounding the core in the cham-
ber permitted water to be injected into the joint under pressure. The injection
pressure and flow outside the joint and the pressure at the exit within the small-
diameter center hole were measured during the test to allow computation of the
joint aperture and hydraulic conductivity. This test provided a direct measurement
of conducting aperture of a specific joint at different normal stress levels. This
procedure, of course, requires a core taken from intervals of jointed rock. Coring
is not recommended for proposed field tests at Libby Dam. However, if suitable
core samples of rock near the lines of drain holes are available from the original
foundation investigation drilling, laboratory joint pressure tests could be per-
formed on selected joints for comparison of joint characteristics calculated from
field tests.
The third step in modeling of uplift pressure distribution and uncertainty is to

construct the numeric flow model to conform to the restrictions imposed by the
field data and the geologic conceptual model.

The fourth step is to run the model and compute uplift pressures.
7.2.2 Additional research exercise for existing dams
Chapter 4 of this report documented a review of existing Corps dams for

assessment of quality of data applicable to research in uplift prediction and
modeling. Libby Dam was selected for case history investigation because of the
good quality of instrumentation and geologic data available. An exercise was
conducted in Chapter 5 to extract hydraulic data from Libby Dam’s foundation
investigation records. Chapter 4 reported that Green Peter Dam had only limited
instrumentation data, but substantial geologic information on stratigraphy, geo-
logic structure, and discontinuities. Green Peter offers an example of a concrete
dam founded on a complex assemblage of jointed, faulted, and sheared inter-
bedded volcanic rocks. A rigorous surface and subsurface foundation mapping
and investigation program provided detailed and comprehensive information
about the geology and engineering characteristics of the foundation rock mass.
Core logging of foundation borings was detailed and stressed attention to the
numbers, locations, and characteristics of discontinuities in the rock mass. While
no data on foundation pressure testing were provided in the foundation report,
pressure testing apparently was performed as part of the foundation grouting,
drainage, exploration, and instrumentation program.
The authors recommend that a case history exercise similar to that described
for Libby Dam in Chapter 5 of this report be conducted for Green Peter Dam. The
goal is to construct a conceptual geologic/hydrologic cross-section transverse to
the dam axis beneath a monolith selected to provide the best data for a potential
uplift numeric model. Any available pressure test data would be used to derive
joint aperture and hydraulic conductivity values, in a manner similar to the exer-
cise for Libby Dam. The development of another case history, which requires
careful scrutiny of available foundation data, would reinforce or improve conclu-
sions drawn from the Libby Dam exercise concerning the kinds and degree of
geologic uncertainty in uplift modeling and prediction.

References
Bandis, S. C. (1980). “Experimental studies of scale effects on shear strength and

deformation of rock joints,” Ph.D. diss., University of Leeds, Department of
Earth Sciences.
Banks, D. C. (1977). “Borehole photography analysis, Teton Dam,” Letter Report,

U.S. Army Engineer Waterways Experiment Station, Vicksburg, MS.
Barton, N. (1973). “Review of a new shear strength criterion for rock joints,”
Engineering Geology 7, 287-332.
__________. (1982). “Modelling rock joint behavior from in situ block tests:
Implications for nuclear waste repository design,” Technical Report,
TerraTek, Inc., Salt Lake City, UT, 65-66.
Barton, N., and Choubey, V. (1977). “The shear strength of rock joints in theory
and practice,” Rock Mechanics 10(1-2), 1-54.
Barton, N., Bandis, S., and Bakhtar, K. (1985). “Strength, deformation and
conductivity coupling of rock joints,” International Journal of Rock
Mechanics, Mining Science, and Geomechanics Abstracts 22(3), 121-140.
Bellier, J. (1977). “The Malpasset Dam” (presented by Pierre Londe). The

evaluation of dam safety, Engineering Foundation Conference proceedings,
Pacific Grove, California, November 28-December 3, 1976. American
Society of Engineers, New York.
Bennett, R. D., and Anderson, R. F. (1982). “New pressure test for determining
coefficient of hydraulic conductivity of rock masses,” Technical Report
GL-82-3, U.S. Army Waterways Experiment Station, Vicksburg, MS.
Bock, R. W., Harber, W. G., and Arai, M. (1974). “Problems encountered in

construction of dam foundations.” Foundations for dams, an Engineering
Foundation conference, Pacific Grove, California, March 17-21, 1974.
American Society of Civil Engineers, New York.
Brekke, T. L., and Howard, T. R. (1972). “Stability problems caused by seams

and faults.” Proceedings, North American Rapid Excavation and Tunneling
Conference, Chicago, IL. American Institute of Mining, Metallurgical, and
Petroleum Engineers, New York, 1, 25-64.
176 References
Davis, S. N., and DeWiest, R. J. M. (1966). Hydrogeology. John Wiley & Sons,
New York.
Deere, D. U. (1964). “Technical description of rock cores for engineering

purposes,” Rock Mechanics and Engineering Geology 1(1), 17-22.
__________. (1981). “Engineering geology for concrete dam foundations.”

Recent developments in geotechnical engineering for hydro projects:
Embankment dam instrumentation performance, engineering geology
aspects, rock mechanics studies, Proceedings sponsored by the Geotechnical
Engineering Division, ASCE International Convention, New York City,
May 11 and 12, 1981. Fred H. Kulhawy, ed., American Society of Civil
Engineers, New York, 166-176.
Ebeling, R. M., and Pace, M. E. (1996a). “Uplift pressures resulting from flow
along tapered rock joints,” The REMR Bulletin, February, 13(1), U.S. Army
Engineer Waterways Experiment Station, Vicksburg, MS.
__________. (1996b). “Variation in uplift pressures with changes in loadings

along a single rock joint below a gravity dam,” The REMR Bulletin, February,
13(1), U.S. Army Engineer Waterways Experiment Station, Vicksburg, MS.
Ebeling, R. M., Pace, M. E., and Morrison, E. E., Jr. (1997). “Evaluating the
stability of existing massive concrete gravity structures founded on rock,”
Technical Report REMR-CS-54, U.S. Army Engineer Waterways Experiment
Station, Vicksburg, MS.
Ebeling, R. M., Nuss, L. K., Tracy, F. T., and Brand, B. (2000). “Evaluation and
comparison of stability analysis and uplift criteria for concrete gravity dams
by three Federal agencies,” Technical Report ERDC/ITL TR-00-1, U.S. Army
Engineer Research and Development Center, Vicksburg, MS.
Ebeling, R. M., Wahl, R. E., and Pace, M. E. (1997). “Analysis of flow and pore
pressures within rock foundations using steady state seepage models,” Draft
Report.
Freeze, A. R., and Cherry, J. A. (1979). Groundwater. Prentice-Hall, Inc.,

Englewood Cliffs, NJ.
Fuller, Mossbarger, Scott and May Engineers, Inc. (2000). “Draft report of
subsurface exploration and data interpretation, Bluestone Dam, Hinton, West
Virginia,” Lexington, KY.
Geotechnical Laboratory (1993). “Part II. In Situ Test Methods, E. Determination

of rock mass permeability.” Rock Testing Handbook. U.S. Army Engineer
Waterways Experiment Station, Vicksburg, MS.
Goodman, R. E. (1968). “Effects of joints on the strength of tunnels,” Technical

Report No. 5, U.S. Army Engineer District, Omaha, Omaha, NE.
References 177
Goodman, R. E. (1990). “Rock foundations for dams: A summary of exploration
targets and experience in different rock types,” Keynote Address: Dam
Foundation Engineering Tenth Annual USCOLD Lecture, New Orleans, LA,
March 6-7, 1990. U.S. Committee on Large Dams, Denver, CO.
Grenoble, B. A., and Amadei, B. (1990). “Evaluation of uplift pressure for con-
crete gravity dams founded on jointed rock: Analytical results.” Dam
Foundation Engineering Tenth Annual USCOLD Lecture, New Orleans, LA,
March 6-7, 1990. U.S. Committee on Large Dams, Denver, CO.
Grenoble, B. A., Harris, C. W., Meisenheimer, J. K., and Morris, D. I. (1995).

“Influence of rock joint deformations on uplift pressure in concrete gravity
dam foundations: Field measurements and interpretation.” Fractured and
jointed rock masses, Proceedings of the Conference on Fractured and Jointed
Rock Masses, Lake Tahoe, California, USA, 3-5 June 1992.
A. A. Balkema/Rotterdam/Brookfield.
Headquarters, U.S. Army Corps of Engineers. (1993). “Seepage analysis and

control for dams,” EM 1110-2-1901, Washington, DC.
__________. (1994). “Rock foundations,” EM 1110-1-2908, Washington, DC.
__________. (1995a). “Geophysical exploration for engineering and environ-

mental investigations,” EM 1110-1-1802, Washington, DC.
__________. (1995b). “Gravity dam design,” EM 1110-2-2200, Washington, DC.
__________. (2001). “Geotechnical investigations,” EM 1110-1-1804,

Washington, DC.
Hietanen, A. M. (1962). “Metasomatic metamorphism in western Clearwater

County, Idaho,” U.S. Geological Survey Professional Paper 344-A,
U.S. Government Printing Office, Washington, DC.
Hvorslev, M. J. (1951). “Time lag and soil hydraulic conductivity in ground-water

observations,” Bulletin No. 36, U.S. Army Engineer Waterways Experiment
International Society for Rock Mechanics. (1978). “Suggested methods for the
quantitative description of discontinuities in rock masses,” International
Journal of Rock Mechanics, Mining Science and Geomechanical Abstracts
15, 319-368.
Iwai, K. (1976). “Fundamental studies of fluid flow through a single fracture,“

Ph.D. diss., University of California, Berkeley.
Lee, C.-H., and Farmer, I. (1993). Fluid flow in discontinuous rocks. Chapman &
Hall, London.
Lohman, S. W. (1972). “Ground-water hydraulics,” Geological Survey

Professional Paper 708, U.S. Government Printing Office, Washington, DC.
178 References
Louis, C. A. (1969). “A study of groundwater flow in jointed rock and its
influence on the stability of rock masses,” Rock Mechanics Research Report
No. 10, Imperial College, London.
Murphy, W. L. (1985). “Geotechnical descriptions of rock and rock masses,”

Technical Report GL-85-3, U.S. Army Engineer Waterways Experiment
Pace, M. E., and Ebeling, R. M. (1998). “Interaction of a gravity dam, rock

foundation, and rock joint with uplift pressures,” Dam Engineering IX(3),
265-305.
Pahl, A., Bräuer, V., and Liedtke, L. (1995). “Fracture flow systems and structural
geology.” Fractured and jointed rock masses, Proceedings of the Conference
on Fractured and Jointed Rock Masses, Lake Tahoe, California, USA, 3-5
June 1992. A. A. Balkema/Rotterdam/Brookfield.
Robinson, E. S. (1982). Basic physical geology. John Wiley & Sons, New York.
Schruben, P. G., Arndt, R. E., Bawiec, W. J., and Ambroziak, R. A. (1997).

Geology of the conterminous United States at 1:2,500,000 scale – A digital
representation of the 1974 P.B. King and H.M. Beikman map. U.S. Geo-
logical Survey Digital Data Series DDS-11, Release 2, U.S. Geological
Survey, Map Distribution, Denver, CO.
Schultz, J. R., and Cleaves, A. B. (1955). Geology in engineering. John Wiley &
Sons, New York.
Sears, F. W., and Zemansky, M. W. (1955). University physics. Addison-Wesley,

Inc., Reading, MA, 1031 pp.
Sharp, J. C. (1970). “Fluid flow through fissured media,” Ph.D. diss., Imperial
College, London.
Snow, D. T. (1965). “A parallel plate model of fractured permeability media,”

Ph.D. diss., University of California, Berkeley.
__________. (1968). “Rock fracture spacings, openings, and porosities,” Journal

of the Soil Mechanics and Foundations Division, Proceedings of the
American Society of Civil Engineers, SM1, 73-91, January.
Stone and Webster Engineering Corporation. (1992). “Uplift pressures, shear

strengths, and tensile strengths for stability analysis of concrete gravity dams,”
TR-100345-V1, Volume 1, Project 2917-05, Final Report, August, Denver,
CO, prepared for Electric Power Research Institute (EPRI), Palo Alto, CA.
Todd, D. K. (1980). Groundwater hydrology. 2d ed., John Wiley & Sons, New
York, 535 pp.
U.S. Army Engineer District, Portland. (1969). “Green Peter Dam, Middle
Santiam River, Oregon, foundation report,” Portland, OR.
References 179
U.S. Army Engineer District, Rock Island. (1965). “Red Rock Reservoir, Des
Moines River, Iowa, Dam – Foundation Report, Binder 1 of 4,” U.S. Army
Engineer District, Rock Island, Rock Island, IL.
U.S. Army Engineer District, Seattle. (1979). “Libby Dam, Foundation Report,”
June, 31 pp plus plates.
__________. (1987). “Libby Dam, Lake Koocanusa Project, Kootenai River,

Montana,” Periodic Inspection Report No. 10, Seattle, WA.
__________. (1993). “Libby Dam, Lake Koocanusa Project, Kootenai River,

Montana,” Periodic Inspection Report No. 13, Seattle, WA.
U.S. Army Engineer District, Walla Walla. (1964). “Dworshak Dam and
reservoir, North Fork Clearwater River, Idaho, Design Memorandum No. 6,
main dam, foundation grouting and drainage, and instrumentation,” Walla
Walla, WA.
__________. (1979). “Dworshak Dam and Reservoir, North Fork Clearwater

River, Idaho, foundation report,” Walla Walla, WA.
Zeigler, T. W. (1976). “Determination of rock mass permeability,” Technical

Report S-76-2, U.S. Army Engineer Waterways Experiment Station,
Vicksburg, MS.
180 References
Appendix A
Explanation of Snow’s
Equation for Flow Through
Fractures
Snow (1968) derived expressions describing flow through smooth-walled

fractures and computing permeability, fracture spacing, and porosity. Figure A.1
describes Snow’s conceptual model and nomenclature.
Figure A.1. Snow’s solid of dimensions W, broken by parallel plane fractures

(after Snow 1968)
From Figure A.1, the (laminar) volume discharge between two smooth
parallel plates of opening 2B is
q = –(B2/3) (γw/µw) (2BW)i (A.1)
where i is the hydraulic gradient. The term –(B2/3) is the intrinsic permeability, k,
but B is 1/2e (the fracture aperture of Figure A.1), so k = (1/2e)2/3) = e2/12.
Appendix A Explanation of Snow’s Equation for Flow Through Fractures A1

Appendix B
Derivation of Darcy’s Equation
of Radial Flow to a Borehole
(Equation 2.11)
H = Ho H=0
Q in
Packer
ro
Ke l
r
R
Borehole
Packer
where
H = excess pressure head
Ho = excess pressure head at the well
Q = flow rate
Ke = interval hydraulic conductivity
Appendix B Derivation of Darcy’s Equation of Radial Flow to a Borehole B1

ro = well radius
r = radius of flow
R = radius of influence
l = length of test interval.
Darcy’s law says:
Q=KiA
where
K = hydraulic conductivity
i = hydraulic gradient
A = cross-sectional area through which flow occurs
For radial flow to the well through a cylindrical interval,
Q = Ke • dh/dr • 2πrl
This can be rewritten as:
dr/r = Ke • 2πl/Q • dh
Integrate at the boundary conditions, where H = Ho at the well and H = 0 at the

radius of influence, R (by definition); r = ro at the well, and r = R at H = 0:
r =R h =0
∫r = ro
dr / r = K e 2 πl / Q • ∫
h = Ho
dh
so ln R – ln ro = Ke 2πl/Q • (0 – Ho)
or ln (R/ro) = –Ke 2πl/Q • Ho
and Q = Ke 2πlHo/ln (R/ro)

(the negative sign is dropped by choosing flow away from the borehole to be
positive).
B2 Appendix B Derivation of Darcy’s Equation of Radial Flow to a Borehole

Appendix C
Glossary
Engineering and Hydraulics Terms

Permeability (k): A general term for ability of a soil or rock to transmit fluid
under a pressure gradient. Intrinsic permeability is independent of the fluid
properties, i.e., it is a property of the medium alone. Intrinsic permeability k =
Cd 2, where C is a dimensionless shape factor describing the geometry and other
characteristics of the flow path and d is the average pore size of the medium
(Davis and DeWiest 1966) or the mean grain diameter (Lohman 1972). Unit is L2
or darcy, where 1 darcy = 1.062 E -11 ft2 (Davis and DeWiest 1966) or 0.987
(µm)2.
Hydraulic conductivity (K): Ability of a medium to transmit fluid considering

the properties of the fluid. K = Cd2 γ/µ = k γ/µ. Units are L/T, e.g., ft/day. Also,
K = Q/iA, where i = δh/δl. Actual units for K are L3/L2/T, as ft3 per ft2 per min,
which reduces to units of velocity, L/T. Formerly called coefficient of
permeability or permeability.
Equivalent hydraulic conductivity (Ke): The hydraulic permeability of a zone of

jointed (discontinuous) rock in which flow is considered to occur in a continuum.
In a borehole test section in a jointed rock mass, flow is assumed to occur over the
entire length of the test section, not just through the open joints. Units are those of
hydraulic conductivity. For a pressure test in a zone of jointed rock, Ke =
(Q/2πLH) ln (R/ro). Formerly called equivalent coefficient of permeability.
Equivalent joint hydraulic conductivity (Kj): The equivalent hydraulic con-

ductivity of a single joint determined from pressure or other tests in a zone of
jointed rock. Kj is derived from calculations of conducting aperture e using the
cubic law (Q is proportional to the cube of the joint conducting aperture).
Joint mechanical aperture (E): The actual distance separating the walls of an
open joint. Most mechanical apertures have a degree of roughness, or asperity,
manifested by irregularities or undulations in the joint walls.
Joint conducting aperture (e): The distance between two smooth, parallel plates
that would allow the same flow as a mechanical joint aperture with rough walls.
Conducting aperture e is always smaller than mechanical aperture E except in the
Appendix C Glossary C1
case of a smooth-walled joint. Also called parallel plate aperture or equivalent
joint aperture.
Cubic law: A relationship stating that rate of flow Q is proportional to the cube of
the joint conducting aperture e. Q = (γw/12µw)⋅e3⋅i for a single joint and e3 =
Q/2πNH (12µw/γw) ln (R/ro) for a borehole test interval containing N joints.
Open joint: A joint having a mechanical (true) aperture greater than about
150 microns (0.150 mm).
Tight joint: A joint having a mechanical (true) aperture less than about
150 microns (0.150 mm).
Reynolds number (Re): A dimensionless number expressing the ratio of inertial

to viscous forces in flow. The critical Reynolds number is the number at which
nonlinear flow starts to occur.
Laminar flow: Flow in which the specific discharge is proportional to hydraulic

conductivity.
Turbulent flow: Flow in which the specific discharge is not proportional to

hydraulic conductivity.
Porosity: The volume of void space in a rock or soil per total unit volume.
Viscosity: The internal friction of a fluid (Sears and Zemansky 1962). The
resistance to flow of a fluid. Dynamic viscosity µ is expressed in units of FT/L2,
e.g., dyne sec/cm2 or lbf min/ft2, or in poise. One poise = 1 dyne sec/cm2. The
dynamic viscosity of water at 20 oC is 0.0101 poise. Kinematic viscosity ν is the
ratio of dynamic viscosity to density and is expressed in units of L2/T, as ft2/min.
Kinematic viscosity ν = g µ/γw.
Selected Geologic Terms

Alluvial: Describing sediments carried by and deposited in streams and stream
valleys.
Amphibolite: A metamorphic rock consisting of dark-colored plagioclase

feldspar minerals and having little or no quartz.
Andesite: A dark-colored igneous extrusive rock.
Anhydrite: A sedimentary rock consisting almost wholly of anhydrous calcium

sulfate. An evaporite rock. Also the mineral anhydrite.
Aperture: A measurement of the distance separating the walls of a discontinuity.
Argillite: A weakly metamorphosed mudstone.
C2 Appendix C Glossary
Asperity: A protrusion or protrusions on the walls of a discontinuity that impart
roughness to the surfaces.
Calcite: A common mineral of calcareous rock, such as limestone and some

sandstones. Also found as fillings in discontinuities. Calcium carbonate, CaCO3.
Clastic: Referring to sedimentary rocks consisting of particles of older rock

accumulating through erosion and subsequent deposition.
Columnar jointing: A kind of joint system developed in volcanic lava flows and
characterized by near-vertical, polygonal sets of fractures formed by cooling of the
lava, particularly basalt.
Dike: A relatively thin, tabular, near-vertical igneous intrusion cutting

transversely through older rock.
Dip: The vertical angle between a discontinuity plane and the horizontal.
Discontinuity: A natural planar or curved feature that physically separates a rock

mass. Examples of discontinuities are joints, bedding planes, faults, shears and
shear zones, planes of schistosity, and mineral veins.
Extrusive rock: Those igneous rocks ejected onto the surface of the earth or into
the earth’s atmosphere from below in a molten or gaseous form. Extrusive rocks
are normally characterized by a fine-grained texture.
Fault: A discontinuity along which the opposite walls have moved past each
other.
Feldspar: A common silicate mineral of igneous rocks, and the most common
mineral in the earth’s crust. Light-colored igneous rocks usually contain sodium-
or potassium-rich feldspar (commonly orthoclase); dark-colored igneous rocks
usually contain calcium-rich feldspars (commonly plagioclase).
Foliation: A general term for structures in metamorphic rocks having platy,

layered, or planar aspects.
Gneiss (adj. gneissic): A metamorphic rock consisting of generally parallel bands

of minerals.
Gypsum: A sedimentary rock consisting almost wholly of hydrated calcium

sulfate. Also the mineral gypsum.
Intrusive rock: Those igneous rocks injected into the earth’s crust from below in
a molten form. Intrusive rocks are normally characterized by a coarse-grained
texture.
Joint: A discontinuity along which there has been no movement or displacement

parallel to the plane of the discontinuity.
Joint set: A group of approximately parallel joints.
Joint system: Two or more joint sets in a consistent pattern.
Kaolinization: The replacement or alteration of minerals to the clay mineral

kaolin. A result of weathering or hydrothermal alteration.
Karst: Terrain characterized by sinkholes and closed topographic depressions,

disrupted surface drainage and disappearing streams, solution-enlarged
discontinuities, caves, and underground drainage systems. Karst is most common
in areas underlain by water-soluble rock.
Lacustrine: Sediments deposited in a lake.
Limestone: A sedimentary rock consisting primarily of the mineral calcite

(calcium carbonate).
Massive: A loosely defined term describing a rock deposit having no visible

bedding or bedding planes spaced greater than a specified thickness.
Metasediments: Certain very old sedimentary rocks that have been sufficiently
altered through time under heat and pressure that they have lost their original
texture.
Porphyry: An igneous rock texture characterized by large mineral crystals in a

matrix of fine crystals.
Pyroclastic: Referring to deposits formed by ejection of molten and gaseous

material from a volcanic vent and subsequently deposited by wind, water, and
gravity. Examples of pyroclastic deposits are ash and tuff, agglomerate, volcanic
breccia, and welded tuff.
Quartz: A common mineral of igneous rocks. Silicon dioxide. Also a major

component of many clastic sedimentary rocks such as sandstone.
Schist (adj. schistose): A metamorphic rock with a strongly foliated texture that
can be readily split into thin flakes or slabs along parallel-oriented minerals.
Shear zone: A narrow zone of intense fracturing, sometimes accompanied by a

gouge of broken and altered rock, along a fault.
Slaking: The tendency of some shales, mudstones, or claystones to disintegrate in

the presence of changing moisture conditions.
Slickensided: Pertaining to striations and grooves on the surface of a

discontinuity indicating movement of the walls of the discontinuity past each
other.
Stratigraphy: That branch of geology that describes the sequence of deposition,

relative age, and lithology of sedimentary and volcanic rocks.
C4 Appendix C Glossary
Strike: The direction, with respect to north, of the line of intersection of a
discontinuity plane with the horizontal.
Tectonic: Referring to forces within the earth that cause mountain building,
uplift, structural basins, earthquakes, and displacement of the earth’s crust.
Tuff: A hardened deposit of volcanic ash. A pyroclastic rock.
Weathering: The in-place disintegration, decomposition, or change in physical

properties of rock by natural processes.
Appendix D
Notation
γ Unit weight of fluid

γw Unit weight of water
µ Dynamic viscosity of fluid
µw Dynamic viscosity of water
σ and σn Normal stress component
τ Shear strength, shear stress
ν Kinematic viscosity
φ Angle of internal friction
A Area
B One-half of the conducting joint aperture (Snow 1968)
c Cohesion
C Dimensionless shape factor
d Pore size
Dh Equivalent hydraulic diameter
e Conducting aperture, parallel plate aperture
E Mechanical aperture
Ex Scientific notation, power (x) of 10
F Units of force
g Acceleration due to gravity
h Head
H Excess pressure head
Hg Head produced by height of water in pressure-test flow pipe
i Hydraulic gradient
I.D. Inside diameter
JCS Joint wall compressive strength
JRC Joint roughness coefficient
k Intrinsic permeability
K Hydraulic conductivity
Ke Equivalent hydraulic conductivity
Kj Joint hydraulic conductivity
Kj′ Turbulent fissure hydraulic conductivity
l Length of borehole test section
L Units of length
MSL Mean sea level
N Number of joints
P Pressure
Appendix D Notation D1
Pt Pressure measured at surface gage
Q Volume flow rate
ro Borehole radius
R Radius of influence
Re Reynolds number
Rr Height of surface asperities
RQD Rock quality designation
S Surface roughness index
T Units of time
u Pore, or uplift, pressure
v Mean specific discharge
Vj Joint closure
Vm Maximum joint closure
D2 Appendix D Notation
Form Approved
REPORT DOCUMENTATION PAGE OMB No. 0704-0188
Public reporting burden for this collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources, gathering and maintaining
the data needed, and completing and reviewing this collection of information. Send comments regarding this burden estimate or any other aspect of this collection of information, including suggestions for
reducing this burden to Department of Defense, Washington Headquarters Services, Directorate for Information Operations and Reports (0704-0188), 1215 Jefferson Davis Highway, Suite 1204, Arlington,
VA 22202-4302. Respondents should be aware that notwithstanding any other provision of law, no person shall be subject to any penalty for failing to comply with a collection of information if it does not
display a currently valid OMB control number. PLEASE DO NOT RETURN YOUR FORM TO THE ABOVE ADDRESS.
1. REPORT DATE (DD-MM-YYYY) 2. REPORT TYPE 3. DATES COVERED (From - To)
March 2002 Final report
4. TITLE AND SUBTITLE 5a. CONTRACT NUMBER
Assessment of Geology as it Pertains to Modeling Uplift in Jointed Rock: A Basis for

5b. GRANT NUMBER
Inclusion of Uncertainty in Flow Models
5c. PROGRAM ELEMENT NUMBER
6. AUTHOR(S) 5d. PROJECT NUMBER
William L. Murphy, Robert M. Ebeling, John M. Andersen

5e. TASK NUMBER
5f. WORK UNIT NUMBER

6
7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES) 8. PERFORMING ORGANIZATION REPORT
NUMBER
Geotechnical and Structures Laboratory and Information Technology Laboratory ERDC TR-02-2
9. SPONSORING / MONITORING AGENCY NAME(S) AND ADDRESS(ES) 10. SPONSOR/MONITOR’S ACRONYM(S)

U.S. Army Corps of Engineers
Washington, DC 20314-1000
11. SPONSOR/MONITOR’S REPORT
NUMBER(S)
12. DISTRIBUTION / AVAILABILITY STATEMENT
Approved for public release; distribution is unlimited.
13. SUPPLEMENTARY NOTES
14. ABSTRACT
Uplift is one of the major forces affecting stability of rock-founded concrete dams. Problems occur in determining the magnitude
and distribution of uplift pressures within rock foundations and in extrapolating uplift pressures to reservoir levels above the pool of
record. Uplift is controlled by the flow regime within the rock foundation. The flow regime is a function of the geology. The goal of
research in uplift uncertainty and probabilistic modeling is to develop a methodology, analytical procedures, and software to assess
uplift pressures and forces within rock foundations for use in the assessment of the reliability of rock-founded concrete gravity dams.
The objectives of this study were to identify geological factors affecting the prediction and modeling of flow and the development of
uplift pressures in rock foundations; to identify the kinds of uncertainty in uplift prediction resulting from geological investigations,
particularly in testing and description of rock discontinuities; and to select a case history for assessing the uncertainties associated with
geological and uplift analysis of a large dam. Much uplift prediction uncertainty is caused by insufficient investigation or treatment of
rock discontinuities, or by deformation of discontinuities with resulting changes in discontinuity aperture and in the flow regime caused
by stresses imposed by the dam and reservoir. Tight discontinuities, those with mechanical apertures less than about 250 microns, may
cause a nonlinear response in uplift pressure with rising headwater. Apertures can be estimated from foundation borehole pressure
(Continued)
15. SUBJECT TERMS
See reverse.
16. SECURITY CLASSIFICATION OF: 17. LIMITATION 18. NUMBER 19a. NAME OF RESPONSIBLE
OF ABSTRACT OF PAGES PERSON
a. REPORT b. ABSTRACT c. THIS PAGE 19b. TELEPHONE NUMBER (include
area code)
UNCLASSIFIED UNCLASSIFIED UNCLASSIFIED 204
Standard Form 298 (Rev. 8-98)
Prescribed by ANSI Std. 239.18
14. (Concluded)
tests using the cubic law, which relates flow rate to joint aperture. There is uncertainty in determining the
persistence of aperture and condition of discontinuities through the foundation rock mass. Other geologic factors
contributing to uncertainty in uplift prediction include variation in rock mass hydraulic conductivity, effects of
weathering, stratigraphic complexity, and variable properties inherent in different rock types. Uplift pressures
measured in gauges in Corps dams commonly reflect pressures only in the upper few feet of the foundation and may
not represent uplift pressures developed in deeper discontinuities. Design engineers usually assume a linear pressure
distribution, when actual pressures may vary nonlinearly both temporally and spatially. Calculation of joint aper-
tures and joint hydraulic conductivity from pressure tests conducted at Libby Dam show decreasing aperture and
hydraulic conductivity with depth, similar to studies at other dams. Apertures calculated for Libby Dam were
generally in the range of “open” joints. Responses of installed pressure gauges at Libby Dam were linear over a
10-year monitoring period, also indicative of open joints.
15. (Concluded)
Dam foundations
Foundation geology
Foundation instrumentation
Hydraulic conductivity
Joint properties
Libby Dam
Pore pressure
Pressure tests
Uncertainty
Uplift
Uplift pressure

Assessment of Geology As It Pertains To Modeling Uplift in Jointed Rock: A Basis For Inclusion of Uncertainty in Flow Models

Uploaded by

Copyright:

Available Formats

Assessment of Geology As It Pertains To Modeling Uplift in Jointed Rock: A Basis For Inclusion of Uncertainty in Flow Models

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Assessment of Geology As It Pertains To Modeling Uplift in Jointed Rock: A Basis For Inclusion of Uncertainty in Flow Models

Uploaded by

Copyright:

Available Formats

ERDC TR-02-2

Risk Analysis for Dam Safety Program

Assessment of Geology as it Pertains

Approved for public release; distribution is unlimited.

The findings of this report are not to be construed as an official

PRINTED ON RECYCLED PAPER

Assessment of Geology as it Pertains

Prepared for U.S. Army Corps of Engineers

Preface ................................................................................................................. xii

Conversion Factors, Non-SI to SI Units of Measurement................................... xiii

1.1 Background ...............................................................................................1

2.1 Groundwater Flow and Uplift Pressure in Discontinuous Rock................3

3—Geological Considerations in Uplift Analysis and Foundation

4.1 Libby Dam ..............................................................................................57

5.1 Background ...........................................................................................107

6.1 Calculation of Uplift Pressures from Gauge Readings ..........................149

7—Conclusions and Recommendations .............................................................169

Appendix A: Explanation of Snow's Equation for Flow Through Fractures ......A1

Figure 2.1. Distribution of uplift pressures in dam foundation for

Figure 2.2. Natural rock joint with mechanical aperture E and

Figure 2.3. Distribution of uplift pressure in (a) joint of uniform

Figure 2.4. Nonlinear response of uplift pressure (as head) to rising

Figure 2.6. Variation in uplift pressure with changes in joint

Figure 2.7. Hyperbolic model for joint deformation .....................................12

Figure 2.8. Empirical curves relating mechanical aperture (E),

Figure 3.1. Distribution of major categories of rock in the

Figure 3.2. Volcanic and metamorphic rocks in the United States................28

Figure 3.4. Relations between geologic structure in abutment of

Figure 3.9. Influence of joint length on uplift pressure distribution..............44

Figure 4.2. Libby Dam, upstream gallery longitudinal section .....................60

Figure 4.3. View upstream of part of upstream grout/drainage

Figure 4.8. Section through Dworshak Dam at station 25+40 ......................67

Figure 4.10. Locations of uplift gauges, Dworshak Dam................................70

Figure 4.15. Plan of Green Peter Dam ............................................................76

Figure 4.16. Sections through Green Peter Dam.............................................77

Figure 4.17. Plan of Green Peter Dam showing locations of borings

Figure 4.19. Locations of pressure measurement instruments, Green

Figure 4.20. Plan view of Detroit Dam ...........................................................82

Figure 4.22. Sections through Detroit Dam ....................................................84

Figure 4.24. Details of instrumentation at Detroit Dam ..................................87

Figure 4.25. Plan view of Wolf Creek Dam....................................................88

Figure 4.26. Sections through Wolf Creek Dam .............................................89

Figure 4.28. Uplift cells in Wolf Creek Dam ..................................................91

Figure 4.29. Plan and section of Old Hickory Dam ........................................93

Figure 4.31. Uplift instruments in Old Hickory Dam .....................................95

Figure 4.32. Plan and section of J. Percy Priest Dam......................................97

Figure 4.34. Locations of borings, J. Percy Priest Dam ..................................99

Figure 3.36. Geologic section along axis of J. Percy Priest Dam..................102

Figure 4.38. Plan view of J. Percy Priest Dam showing locations of

Figure 5.2. Mapped discontinuities on foundation floor, monoliths

Figure 5.3. Schematic of borehole pressure tests, Libby Dam ....................114

Figure 5.9. Summary log of boring D-126, monolith 23 area, Libby

Figure 5.11. Summary log of boring D-188, monolith 23 area, Libby

Figure 5.12. Summary log of boring D-278, monolith 23 area, Libby

Figure 5.15. Equivalent hydraulic conductivity, number of joints per

Figure 5.17. Equivalent hydraulic conductivity, number of joints per

Figure 5.18. Equivalent hydraulic conductivity, number of joints per

Figure 5.20. Equivalent hydraulic conductivity, number of joints per

Figure 5.21. Variation in equivalent hydraulic conductivity with