ADVANCED DAM

ENGINEERING
FOR DESIGN,
CONSTRUCTION, AND
REHABILITATION
I

Edited bY

Robert B. Jansen
Consulting Civil Engineer

ffiI VAN NOSTRAND REINHOLD

tultr{ New York
 Copyright O 1988 by Van Nostrand Reinhold

Library of Congress Catalog Card Number Az-ZiOoo

lsBN H42-24397-9

All rights reseNed. No part of this work covered by the copyright hereon
may be reproduced or used in any form or by any means-graphic, electronic,
or mechanical, including photocopying, recording, taping, oi iniormation
storage and retrieval systems-withoul written permission of the publisher.

Printed in the United States of America

Van Nostrand Reinhold

115 Fitth Avenue
New York, New York 10003

Van Nostrand Reinhold lnternational Company Limited

11 New Fetter Lane
London EC4P 4EE, England

Van Nostrand Reinhold

480 La Trobe Street
Melbourne, Victoria 3000, Australia

Macmillan of Canada
Division of Canada Publishing Corporation
164 Commander Boulevard
Agincourt, Ontario M1S 3C7, Canada

16 15 14 13 12 11 10 9 8 7 6 5 4 3 2,1

Llbrary ol Congress Cataloglng-ln-publlcatlon Data

Advanced dam engineering for design, conslruction,

and rehabilitation.

lncludes index.
1. Dams-Design and construclion. l. Jansen,
Robert B., 1922-
TCS0.A33 1988 627'.8 87-21066
lsBN 0-442-24397-S

-,.{
l
,T
.d
 PREFACE
The present state of the art of dam engineering has been
attained by a continuous search for new ideas and methods
while incorporating the lessons of the past. [n the last 20
years particularly there have been major innovations, due
largely to a concerted effort to blend the best of theory and
practice. Accompanying these achievements, there has been
a significant trend toward free interchange among the pro-
fessional disciplines, including open discussion of prob-
lems and their solutions. The inseparable relationships of
hydrology, geology, and seismology to engineering have
been increasingly recognized in this field, where progress
is founded on interdisciplinary cooperation.
This book presents advances in dam engineering that
have been achieved in recent years or are under way. At-
tention is given to practical aspects of design, construction,
operation, and rehabilitation. Case histories are reviewed
to demonstrate principles and procedures of proven value.
Lessnns drawn from, or reinforced by, accidents and fail-
-;ies are examined. The fundamentals of this field are cov-
ered primarily as a means of introducing and facilitating an
understanding of concepts and practices that represent the
leading edge of technology. Both simplified and compre-
hensive procedures are presented. Strong emphasis is
placed on technical aspects, with only general considera-
tion given to administrative, financial, social, legal, envi-
ronmental, and political factors, which, thou_sh important,
are covered in other publications.
The rapid progress in recent times has resulted from the
combined efforts of engineers and associated scientists, as
exemplified by the authorities who have contributed to this
book. These individuals have brought extensive knowledge
to the task. drawn from experience throughout the world.
With the convergence of such distinguished talent, the op-
portunity for accomplishment was substantial. I gratefully
acknowledge the generous cooperation of these writers, and
am indebted also to other persons and organizations that
have allowed reference to their publications; and I have
attempted to acknowledge this obligation in the sections
where the material is used. These courtesies are deeply ap-
preciated.
Special credit is due to C. F. Corns, T. P. Dolen, L. B.
James, R. W. Kramer, T. M. Leps, G. Lombardi, J. Lowe
III, J. M. Raphael, and E. T. Scherich for their work as
reviewers, and to those who served as coordinators,
namely. C. B. Cecilio, C. F. Corns, C. A. Fetzer. L. B.
James, P. C. Knodel, R. W. Kramer, J. D. Lytle, E. T.
Scherich, and G. S. Tarbox.
The participants hope that they have contributed usefully
to improved practices in the engineering of dams.
Robert B. Jansen
 CONTENTS
PREFACE v 12. ROCKFILL DAM OES]GN AND ANALYSIS 368
Thomas M. Itps
1. INTRODUCTION Robert B. lansen 1
13. ROCKFILL DAM CONSTRT.ICTION ANO 388
2. LESSONS FROM NOTABLE EVENTq 8 FOUNDATION TREATMENT Thomas M.
Laurence B. James, Roben B. Jansen, LePs
George A. Kiersch, Thonws M. I-eps
14. ROCKFILL DAM PERFORMANCE AND 409
3. HYDROLOGY Catalino B. Cecilio, 60 REMEDIAL MEASURES Thomas M. I*ps
Llewellyn L. Cross, Arlen D. FeArnan
15. EARTHOUAKE RESPONSE ANALYSIS OF 416
4. GEOLOGY Arthur B. Arnold, Laurence B. 106 CONCRETE DAMS Anil K. Chopra
lames, George A. Kiersch, Alan L. O'Neill
. 16. GRAVITY DAM DESIGN AND ANALYSIS 466
5. SETSMOLOGY Bruce A. Bolt 153 Charles F. Corns, Ernest K. Schrader,
Glenn S. Tarbox
6. MATERIALS Timothy P. Dolen, Claude A. 172
Fetzer, Robert B. Jansen, Paul C. Knodel, 17. ARCH DAM DESIGN AND ANALYSIS 493
Ernest K. Schrader, Ltoyd O. Timbtin,Jr. Howard L. Boggs, Robert B. Jansen, Glenn
S. Tarbox
7. COFFERDAMS Claude A. Fetzer, Edwin 219
Paul Swatek, Jr. 18. CONCRETE DAM CONSTRUCTION AND 540
FOUNOATION TREATMENT Ernest K.
8. EARTHOUAKE RESPONSE ANALYSIS OF 239 Schrader, william F. swiger
EMBANKMENTS, Jcmes Michael Duncan,
I. M. Idriss 19. CONCRETE DAM PERFORMANCE AND 578
REMEDIAL MEASURES Charles F. Corns,
9. EARTHFILL DAM DESIGN AND ANALYSIS 256 RobeN B. Jansen, GiOVANNi I.OMbATdi
Robert B. Jansen, Richard W. Kramer, John
Lowe III, Steve J. Poulos 20. SPILLWAY DESIGN ANO CONSTRUCTION 609
John Lowe III, Nelson L. de S. Pinto,
tO. EARTHFTLL OAM CONSTRUCTION AND 321 Richard P. Regan, E. Thonws Scherich,
FOUNDATION TREATMENT Claude A. Ernest K. Schrader
Fetzer, Richard W. Kramer, William F.
swiger 21. SPILLWAY PERFORMANCE ANO 647
REMEDIAL MEASURES PhiliP H. Burgi,
11. EARTHFTLL DAM PERFORMANCE AND 354 Edward W. Gray, Peter G. Grey, David L.
REMEDTAL MEASURES Claude A. Fetzer, Hinchlifi, Nelson L. de S. Pinto, Richard P.
Richard ll'. Kramer, William F. Swiger, Regan, E. Thomas Scherich, Ernest K.
Jack G. Wutff Schrader
Ylll Contonts

22. OUTLET DESIGN AND CONSTRUCTION 686 25. INSTRUMENTATION Lynn R. Carpenter, J. 7S1
Roben B. Jansen, Richard P. Regan, E. David Lytle, Dewayne L. Misterek, Bntce C.
Thomas Scherich Murray, Jerome M. Raphael

23. OUTLET PERFORMANCE AND REMEDIAL 7O4

MEASURES Philip H. Burgi, Roben B.
Jansen, Richard p. Regan, E. Thomas 26. suRvElLLANcE Lloyd A. Duscha, Robert 777
Scheich, Ernest K. Schrader B' Jansen

24. RESERVOIRS l-aurence B. James, George lZ2

A. Kiersch TNDEX 799
 ADVANCED DAM
ENGINEERING
FOR DESIGN,
CONSTRUCTION, AND
REHABILITATION
 INTRODUCTION
RoseRr B. JexsEN
Consulting Civil Engineer
Bellingham, Washington
HISTORY
Since its beginning in a distant era. the engineering of dams
has evolved from crude works to increasingly complex sys-
tems. Ancient dam construction was a tentative art drawn
from simple experimentation. Over many centuries. the art
was joined gradually with science.
The early dam builden made practical use of alluvium
that was readily available along streams. They created pre-
carious embankments, easily erased by floods. As a con-
sequence, earthfills inspired little confidence for thousands
of years.
Dams made of rock were favored in many parts of the
world. In the Wadi el-Garawi near Helwan in Egypt are
remnants of Sadd el-Kafara Dam. which dates from as early
as 2600 B.c. to 2900 s.c. It was a 46-ft-high (I4-meter)
rockfill with cut-stone masonry faces and a core of rubble
and gravel. The central section is assumed to have been
destroyed by flood soon after, or in the final stage of. con-
struction. The core materials were not well protected from
erosion. Insufficient provision for spill has been responsi-
ble for many dam failures throughout history.
As the centuries unfolded, craftsmanship improved. The
Romans constructed many dams of very durable mortared
masonry. Some of them are still in service, their large spill-
ways attesting to the Roman understanding of engineering
fundamentals. Some historians believe that the Romans
were the originators of the arch dam. An impressive struc-
ture of this type was built on a tributary of the Khabur
River near Daras on the Turkish-Syrian border. An ac-
count of this construction by Byzantine engineer Chryses
t'rom Alexandria was documented in about e.o. 560. The
steep abutments were notched to ensure anchorage of the
arch, and gates reportedly were installed for tlood control.
Although the project incorporated such significant features,
the old records shed little light on its structural dimensions.
A notable dam of the Mongol period in Penia was built
in a narrow gorge on the Kebar River. about 15 miles (2'l
km) south of the town of Qum. Dating from approximately
1300 e.o., it is believed to be the oldest surviving arch
dam, and is about 85 ft (26 meters) high and 180 ft (55
meten) long. The dam is composed of cemented rubble
masonry with mortared stone block facing. The arch was
keyed into the canyon walls.
The engineering of buttress dams was advanced by the
construction of Albuera de Feria Dam (also known as Al-
mendralejo Dam) near Badajoz. Spain. in 1717. h is a rub-
ble-masonry structure. 77 ft (23.5 meters) high and 558 ft
(170 meten) long, with buttresses supponing its down-
stream face.
Design concepts for stone dams were conveyed from
Spain to the Americas. In Mexico. the rubble-masonry but-
tressed type was adopted at many sites.
The first multiple-arch dam is considered to be the nra-
sonry structure called Meer Allum. erected near Hydera-
bad. India, in about 1800. It is approximately 40 ft (12
meters) high and 2500 ft (762 meters) long.
Until about t850. there were few rational criteria for the
design of dams. Failure of Puentes Dam on the Rio Guad-
alentin in Spain in 1802 demonstrated the inadequacy of
some empirical methods. This 164-ft-high (5O-meter) ma-
sonry gravity structure was intended to be based entirely
on rock. but discovery of deep alluvium in the river chan-
nel led to use of a system of wood piles and bracing in the
foundation at that section. After I I years of service, the
undcrpinning blew out, lcaving thc dam spanning a large
breach.
During thc ninetecnth century, notabte progress was
made in European gravity dam engincering. In 1853.
French engineer M. cle Sazilly advocated that prcssures
within a dam be kept within specified limits. and that di-
mensions be set to prevent sliding. The concept of main-
 2 ADVANCED DAM ENGINEERING FOR DEStGN, CONSTRUCTION, AND REHAEILITAT]ON
taining the resultant of forces within the middle third of
each horizontal plane was expounded about 25 years later
by W. J. M. Rankine of England. The conceprs advanced
by M. de Sazilly and Rankine led toward logical analysis
of dams.
While rhe middle-third criterion was generally accepted
as ensuring overturning resistance of moderately loaded
gravity dams, engineers began to recognize the importance
of uplift and sliding. A drainage system to limit uplift was
adopted in the design of Vyrnwy Dam in England, con-
structed in the period 1882-90.
Important advances have been made in concrete tech-
nology for dams since the early 1900s, including accp-
rately controlled mixes, vibration, special cements, poz-
zolans, aggregate processing, air entrainment, cooling, and
admixtures. Progress in concrete dam design and construc-
tion is exemplified by monumental structures such as the
935-ft-high (285-meter) Grand Dixence Dam in Switzer-
land (1962).
Development of embankment construction practices was
accelerated in the nineteenth century as machinery replaced
animals for transport and compaction. The transition from
horse-drawn wagons and graders, smooth rollers, and har-
rows was completed soon after 1900 as heavy equipment,
including sheepsfoot compactors and vehicles with crawler
tracks. came into general use. By 1940, Iarge special
equipment for moving and compacting earth and rock had
been developed. In the 1950s and 1960s, the vibrating roller
won wide acceptance, first for compacting sands and grav-
els and then on rockfills. It had an important influence in
the adoption of compacted rockfill dams in preference to
dumped fills.
With the substantial increase in construction capability,
there was a trend toward larger embankments, which was
strengthened by general recognition oftheir inherent ability
to accommodate a broader range of site conditions than
would be suitable for more rigid dams.
Rockfill construction in America gained impetus at Salt
Springs Dam in California (1931), which set a new record
height of 328 ft (100 merers). This is a dumped fill with a
concrete face, a prevalent type until about the middle of
the current century. Some dumped rockfills with upstream
slabs experienced substantial settlement and leakage. Their
popularity declined as engineers turned to earth-cored
rockfills in the 1940s and concrete-faced compacted rock-
fills in the 1960s. More recent rockfills that have estab-
Iished world marks include Esmeralda (Chivor) Dam in
Colombia (1975),777 ft (237 meters), and Chicoasdn Dam
in Mexico (1980), 856 ft (261 meters). Both have earth
cores.
The sizes of earthlills also have increased in rapid
succession. Among those completed in the last 20 years are
the 754-ft (230-meter) Oroville (United States, l96g), the
794-ft (242-meter) Mica (Canada, 1973), and the 9g4-ft
(300-meter) Nurek (Soviet Union, l9g0). These eanhfi[s
have outer zones of gravel.
Until the twentieth century most of the knowledge ap-
plied in the design of embankments was empirical. Their
ability to withstand imposed florces was not easily deter_
mined; limits were judged largely on the basis of accidents
and failures. In contrast, analytical methods for concrete
and masonry dams were considered relatively dependable,
although derived from simple concepts. Gravity dams. for
example. were dimensioned amply to ensure resistance to
sliding and overtuming. The calculated tensile stresses in
such structures were usually minimal.
Numerical analysis of both concrete and embankment
dams has improved substantially in the past 50 years. The
sciences have gained a prominent role, enabling engineers
to develop large, economical, reliable structures while cop-
ing with the reality of fewer ideal sites. The continued
blending of practical and scientific approaches provides a
sound basis for further advances.
BASIC CONSIDERATIONS
The boundaries ofdam engineering have been expanded to
a large extent by applying lessons from projects that have
failed. Most problems have been attributable to simple
causes, often related to misjudgment of foundation capa_
bility. The natural formations at sites of dams and reser_
voirs usually have widely ranging abilities to preclude or
convey water. Each damsite has its individual characteris-
tics, which require special consideration.
Besides assessing the demands of more difficult sites for
future dams, we must be watchful of those engineered in
the past. Some of the most critical variables relate to foun-
dation perforrnance, both structural and hydraulic. The
safety of a dam is inseparable from the capability of its
foundation. Yet, exploration made for design purposes may
not disclose all geological conditions at a site so that the
most vulnerable zones can be defined. The extent of defi-
ciencies may not become apparent until the reservoir has
been operated through several cycles.
Effective dam engineering requires respect for the de-
structive potential of uncontrolled water. To ensure that its
force is subdued, it must be excluded or confined to safe
passages. A dam will alter the natural equilibrium at its
site. As water is impounded, a new regime of seepage will
develop in the barriers that form the reservoir. The char-
acteristics of materials in the dam and in its foundation will
change. causing a transition in the capabilities of the struc-
ture. The capacity of the natural formation to carry the su-
perimposed loads may depend to an important extent upon
the effects of seepage, including erosion, solution, and in-
terstitial pressures.
Evaluation of the behavior of dams calls for knowledge

of a wide range of technologies and practices employed by
differcnt engineers over many yean. For essential conti-
nuity. designers must be involved in construction and op-
eration. They have to share their understanding of how the
project was intended to function, and they must assist in
judging actual conditions against design expectations. The
engineers responsible for design and construction of a dam
will develop special insights into its foundation and its ma-
terials. They should know where the site and the structure
have potential problems. To facilitate future analysis. their
information and its limits should be well rccorded. Without
such documentation. those entrusted with care of the dam
as it ages may be at a disadvantage. A dam that has out-
lived its builders may be an enigma to others in its later
years.
TYPES OF DAMS
Dams are classified usually in terms of materials and form.
Common types are homogeneous or zoned earthfills: rock-
fills with earth core or concrete face; and concrete dams
that depend upon gravity, arch, or buttress resistance.
Structures supported by buttresses are further categorized
as flat-slab (Ambunen), multiple-arch, or massive-head.
Some dams are composites of various materials, including
earthfill. rockfill, masonry, and concrete. A few have tim-
ber, asphaltic, or synthetic membranes.
Topography and geology are primary factors in wei,ehing
the comparative merits of dam types. These interrelated
characteristics of the site influence the loading distribution
on the foundation and the seepage patterns through the res-
ervoir margins. Stress concentrations may occur in a dam
where there are sharp differences in the deformability or
the shape of its foundation.
Earth embankments particularly may be damaged by dis-
tortions at critical points. Differential settlement may be
severe at steep abutments and at structural interfaces where
effective compaction is difficult to obtain. At these loca-
tions. deformation of the fill may open dangerous paths of
seepage. For this reason, there have been many failures
along outlet conduits.
Although properly constructed embankments are able to
accommodate substantial movement. they have relatively
poor resistance to overflow; so their freeboard and associ-
ated spillway capacity must be determined conservatively.
In contrast, most concrete dams can withstand overtopping
for at least several hours. The key to their safety may be
the resistance of the foundation to impact of the spill.
Arch dams can carry large loads, but thcir integrity dc-
pends inherently on the strength of the abutments. Failurc
may be caused by rock deterioration or by shearing under
water pressures. Weakening of arch support also ntay be
triggered by foundation erosion.
TNTROOUCTIOI{ 3
Gravity tlams arc noted for durability. Bc'cause of their
large masses. they can survive considerable weathering and
site dchciencies. However. some have failed wherc foun-
dation elements were susceptiblc to sliding. A few but-
tresscd dams also have shown this tendcncy.
Each kind of dam has distinctivc fcatures. but thc'y
should not be viewed as invariablc. Certainly, when site
specifics are considered. a given type of dam will show
characteristic merits or disadvantages. Nonetheless. eval-
uation in terms of stereotypes may be mislcading. A rtr*k-
fill. for example, is generally regarded as capable of en-
during high rates of leakage for extended periods. Although
this is true in many cases. some rockfills contain substan-
tial concentrations of fine materials that mi-sht be suscc'p-
tible to erosion, especially those constructed by dumping
in high lifts.
ELEMENTS OF DAM ENGINEERING
Engineers need periodically to review and adjust their per-
spectives. Established practices must be reexamined as re-
.sources become more limited. and as the demands upon
them increase. New technical knowledge has to be shared
fully. Of course, old methods should not be lightly aban-
doned. nor should all that is new be embraced. Each must
stand the test of current circumstances.
Despite the important advances attained in modern times.
dam engineering has many acknowledged limitations. The
extent and the character of risk are not totally predictable
in every case. A primary goal is to reduce uncertainties
through investigation and preventive and remedial effon-
In contemplating a project at its beginning. one must ex-
amine the setting thoroughly to foresee problems. The en-
gineers and the allied scientists. including geologists. seis-
mologists, and hydrologists. must pool their knowledge and
ideas so that the appropriate type of dam is selected and
designed to withstand postulated loadings. This collabo-
ration has to be a continuin-e process. Unknown problems
may remain as the works are put into service. Risk must
be recognized as a variable dependent upon changing struc-
tural and site conditions. as well as developments in the
area of potential impact. The rational approach to dam en-
gineering requires anticipation of the full array of circum-
stances that might occur. drawing from experience at dams
where tolerable limits have been exceeded. In weighing
potcntial weaknesscs, the engineer must be aware of the
inaccuracics of santpling techniqucs in ficld and laboratory
work. Results may vary ovcr a broad range. Natural ma-
tcrials may have average Propcrtics that mcet requirc-
ments, but may bc unacccptable whcn judged by their vari-
ations.
Nonunifonnity of ntatcrials is a source of problems in
some older concrete dams. [t has been related to many de-
 4 ADVANCED DAM ENGINEERING FOR OESIGN, CONSTRUCTION, AND
REHABITITATTON
ficiencies of earthfills. Embankment zones will not be com-
pletely homogeneous despite strict construction supervi-
sion. The designer may assume uniformity to facilitate
analysis, but the materials in the dam will be variable. Bor-
row areas and quarries may provide soils and rock that dif_
fer from what was expected. even though the specifications
and their enforcement are exacting. A permeable layer
placed within a zone designed to be impervious may con_
stitute a conduit through the embankment. An impermeable
layer intended to be pervious may preclude necessary
drainage. Omission of a filter between fine and coarse zones
could result in migration of particles from one zone to the
other. Moisture, density, and gradation of the fill materials
will not be uniform. Also, consolidation can cause perrne_
ability to vary considerably from crest to base. To cope
with these conditions, the design must provide successive
lines of defense. Filters and drains may prove vital to the
control of seepage conditions that calculations do not pre_
dict. Escaping water may cause great damage on the un_
denide of an earthfill, and a common grout cuftain alone
may not adequately protect materials at the foundation con_
tact. It could be least effective at its top, where grouting
pressures typically would have been low. The fill should
be isolated from detrimental underflows by measures such
as consolidation grouting, slush grouting, concrete dental
work, or placement of filters.
Defensive engineering requires incorporation of such
protective elements to counter every adverse condition.
Each project demands its own measures to meet the haz_
ards peculiar to the site. Even after extensive exploration,
testing, and analysis, the designer must be mindful of the
uncertainties, arid should call for enough treatment to com-
pensate for them.
In the design and analysis of dams, concepts drawn from
reasonable assumptions may be of greater value than cal_
culations based upon uncertain data. Because dam con_
struction requires the use of materials and foundations that
are inherently nonuniform, the engineer must maintain a
realistic perspective on their capabilities and limitations
while applying the available scientific methods. Sophisti-
cated devices and techniques may give misleading impres_
sions of precision. Theories are founded on premises that
may be valid only within limited ranges, and nearly all of
them must be tested by practical experience. Engineers
must exercise their own judgment to the fullest in the final
analysis, drawing from its sources in theory and practice.
RECENT PROGRESS
In the past, when sites were favorable and resources plen_
tiful, the shortcomings of science could be overcome by
generous budgets. Massive dams stand today as monu_
ments to that era and to the greatncss of its engincers. Many
of those works may outlive the prcsent civilization. surviv_
ing probable maximum floods, maximum earthquakes. and
an inevitable measure of neglect. They were sound invest-
ments in their time.
Today, greater emphasis must be placed upon the eco_
nomical use of resources. Design criteria must be scruri-
nized to elintinate excesses. [t is necessary to determine
levels of acceptable risk for different sets of conditions.
Attention must also be given to the justification of any dil_
ferent criteria for old dams and new dams.
Procedures for the engineerin,e of dams have evolved un-
der the authorship of an array of distinguished professional
talent. Revisions or exceptions should not be advocated
without compelling reason. Nonetheless. the need for fur-
ther improvement is clear. This applies particularly to older
structures. Practical and efficient methods must be found
to raise the capabilities of dams within the restraints im_
posed on them. A resource that must be tapped to the full-
est is the ability of engineers and their professional asso_
ciates to innovate in ways that will be effective. This is the
essential challenge.
The cu rrent accelerated tech nolog ical progress cont rasts
sharply with the gradual evolution of theories and practices
in earlier centuries.
Some noteworthy contributions to the advancement of
dam engineering in recent times are described below.
New Perspectives on Seismology
Important progress has been made in mathematical mod_
elin-g of seismic fields near their sources. Improved pro_
cedures are used for estimating attenuation of accelerations
and for developing time histories and spectra.
lmproved Methods for Foundation Treatment
There is increasing reco_enition of the value of grouting with
thick stable mixes and high pressures. New methods are
available for improving cement used in grouting, a primary
objective being to attain a fineness that facilitates penetra_
tion of fissures. The Wet Milled Cement Method, devel-
oped in Japan, enables the refinement of common cemenr
at the site. A mill with high-speed rotary disks is installed
in the grouting system between mixers. The reground ce_
ment yields a stable mix with improved ability to penerrate.
An unusually extensive grout curtain of the bathtub type
was established at El Caj6n Dam, a 76g-ft-high (234-me_
ter) high arch dam in Honduras. The massive and moder-
ately karstificd limestone foundation was injected with
100,000 metric tons of cemenr, in one thick stable mix,
 4 AOVANC€D DAM ENGITVEERING FOR DESIGN. CONSTRUCTTON,
AND REHABILITATION
ficiencies of earthfills. Embankment zones will not be com_
pletely homogeneous despite strict construction supervi-
sion. The designer may assume uniformity to facititate
analysis, but the materials in the dam will be variable. Bor-
row areas and quarries may provide soils and rock that dif-
fer from what was expected, even though the specifications
and their enforcement are exacting. A permeable layer
placed within a zone designed to be impervious may con-
stitute a conduit through the embankment. An impermeable
layer intended to be pervious may preclude necessary
drainage. Omission of a filter between fine and coarse zones
could result in migration of particles from one zone to the
other. Moisture, density, and gradation of the fill materials
will not be uniform. Also, consolidation can cause perrne_
ability to vary considerably from crest to base. To cope
with these conditions, the design must provide successive
lines of defense. Filters and drains may prove vital to the
control of seepage conditions that calculations do not pre_
dict. Escaping water may cause great damage on the un_
denide of an earthfill, and a common grout curtain alone
may not adequately protect materials at the foundation con-
tact. It could be least effective at its top, where grouting
pressures typically would have been low. The fill should
be isolated from detrimental underflows by measures such
as consolidation grouting, slush grouting, concrete dental
work, or placement of filten.
Defensive engineering requires incorporation of such
protective elements to counter every adverse condition.
Each project demands its own measures to meet the haz_
ards peculiar to the site. Even after extensive exploration,
testing, and analysis, the designer must be mindful of the
uncertainties, and should call for enough treatment to com-
pensate for them.
In the design and analysis of dams, concepts drawn from
reasonable assumptions may be of greater value than cal_
culations based upon uncertain data. Because dam con_
struction requires the use of materials and foundations that
are inherently nonuniform, the engineer must maintain a
realistic perspective on their capabilities and limitations
while applying the available scientific methods. Sophisti_
cated devices and techniques may give misleading impres_
sions of precision. Theories are founded on premises that
may be valid only within limited ranges, and nearly all of
them must be tested by practical experience. Engineers
must exercise their own judgment to the fullest in the final
analysis, drawing from its sources in theory and practice.
RECENT PROGRESS
In the past, when sites were favorable and resources plen_
tiful, the shortcomings of science could be overcome by
generous budgets. Massive dams stand today as monu_
ments to that era and to the greatncss of its engineers. Many
of those works may outlive the present civilization. surviv-
ing probable maximum floods, maximum eanhquakes. and
an inevitable measure of neglect. They were sound invest-
ments in their time.
Today, greater emphasis must be placed upon the eco-
nomical use of resources. Design criteria must be scruti_
nized to eliminate excesses. It is necessary to determine
levels of acceptable risk for different sets of conditions.
Attention must also be given to the justification o[ any dif_
ferent criteria for old dams and new dams.
Procedures for the engineering ofdams have evolved un-
der the authonhip of an array of distinguished professional
talent. Revisions or exceptions should not be advocated
without compelling reason. Nonetheless. the need for fur-
ther improvement is clear. This applies particutarly to older
structures. Practical and efficient methods must be found
to raise the capabilities of dams within the resrrainrs im-
posed on them. A resource that must be tapped to the tull_
est is the ability of engineers and their professional asso-
ciates to innovate in ways that will be effective. This is the
essential challenge.
The current accelerated technological progress contrasts
sharply with the gradual evolution of theories and practices
in earlier centuries.
Some noteworthy contributions to the advancement of
dam engineering in recent times are described below.
New Perspectives on Seismology
Important progress has been made in mathematical mod_
eling of seismic fields near their sources. Improved pro_
cedures are used for estimating attenuation of accelerations
and for developing time histories and spectra.
lmproved Methods for Foundation Treatment
There is increasing reco-enition of the value of grouting with
thick stable mixes and high pressures. New methods are
available for improving cement used in grouting, a primary
objective being to attain a fineness that facilitates penetra_
tion of fissures. The Wet Milled Cement Method, devel-
oped in Japan. enables the refinement of common cemenr
at the site. A mill with high-speed rorary disks is installed
in the grouting system between mixers. The reground ce_
ment yields a stable mix with improved ability to penetrate.
An unusually extensive grout curtain of the bathtub type
was established at El Caj6n Dam, a 76g-ft-high (234_me_
ter) high arch dam in Honduras. The massive and moder-
ately karstified lirnestonc foundation was injected with
100,000 metric tons of cement, in one thick stable mix,

using pressures up to 50 bar. The approximate surface area
of the cunain is 634,000 yd: 153O.OOO mr;.
In the construction of Ouchi Dam. a rockfill in Japan.
grouting of the foundation was accomplished by a tully au-
tomatic system consisting of a central plant. secondary
plants. and injection points. The central plant containcd
mixers. stor?ge and conveyance facilities for undiluted
grout. and a control room. The undiluted grout was con-
veyed to agitators and injection pumps in the secondary
plants. along with water to dilute it to a controlled con-
sistency. Injection rates and grout pressures and thick-
nesses were registered and governed automatically at the
computer station in the control room. This system enabled
simultaneous operation of a complex of grout pumps.
Permeable foundation soils can be grouted deeply by
using hi-eh-pressure pumps, which may be provided with
automatic recording. Effective -erouts include cement-ben-
tonite, deflocculated bentonite. and silicate-based mix.
Where a grout curtain would be infeasiblb or of doubtful
value. there is a trend toward adoption of vertical walls and
diaphragms as cutoffs. using clay-cement. conventional
concrete. or plastic concrete as trench backfill. The more
deformable wall materials are generally favored in alluvial
foundations. particularly where seismic loading might be
severe. Notable advances have been made in techniques
and equipment for installing these walls.
Progress in Soil Mechanics
Site exploration has been enhanced by advanced -eeophys-
ical methods and by use of in-site soil-testing devices.
Testing and treatment of dispersive soils have'been im-
proved. The pinhole test was developed for identification
of these soils, which have been shown to be highly suscep-
tible to colloidal erosion. High pore-water sodium was
confirmed to be a principal characteristic of most disper-
sive soils. Some older embankments containin-g dispersive
materials have been treated successfully with additives
(alum, gypsum, and hydrated lime) to improve their resis-
tance to erosion.
A better understanding of the behavior of soils during
cyclic loading has led to more rational analysis of embank-
ment stability.
Criteria have been established for identification of ma-
terials susceptible to liquefaction. New approaches are used
in the dynamic analysis of embankments or foundations
containing these soils.
Extensive studies have providcd ncw insights into thc
mechanisms of hydraulic fracturing, internal erosion, sub-
sidence. and soil collapse.
Research and field experience have expandcd knowlcdge
of the strength and deformation characteristics of earth-
INTRODUCTION 5
rock mixtures. Effcctive mcthods have bccn dcveltlpcd for
placement and compaction contrul of thcsc matcrials.
Reliable critcria are applied in the dcsign of tilten and
drains for embankmcnts. Seismic etlccts on corc and liltcr
materials have bccome morc predictable.
Computerized Seepage AnalYsis
The design of scepagc control measurcs tbr an embank-
ment and for its foundation has been optimizcd by rcfincd
computer-based procedurcs. enabling panmetric studies of
permeabilities and prediction of pore Prcssures tbr use in
stability analyses. Scepage can be analyzed using elcctric-
analogy models or finite-element techniques. Sat'ety factors
are introduced to allow for the variability of natural mate-
rials and for eventual decline in efficiency of the seepage
control system.
lmproved Stability Analysis of Embankments
Finite-element models have been applied for the dynamic
response analysis of embankment dams.
Simplified methods are now available for estimating em-
bankment deformation.
Synthetic Membranes
Geomembranes and -geotextiles are gaining acceptance. un-
der certain conditions. for filtering. drainage. and soil re-
intorcement.
Methods of Earth Reinforcement
Earth reinforcement techniques have found usetul appli-
cations in the construction and the rehabilitation of em-
bankment dams. Reinforced earth walls have been added
for heightening earthfi lls.
Construction on Soil Foundations
Dynamic compaction is used increasingly to remedy defec-
tive soil foundations. [n construction of Jebba lr{ain Dam.
a zoned rockfill in Nigeria. foundation densification on a
massive scale was accomplished by deep blasting and vi-
brocompaction. These methods were effective in improv-
ing alluvial sands that in their natural state wcre locally
loose at various depths.
Tarbcla Dant in Pakistan set a ncw record as the largcst
enrbankmcnt darn in thc world. lnnovative nlcthods were
used to copc with challcnging conditions at the site on the
Inclus River. The ntain embanknlent was founded on allu-
viunr as deep as 700 li (2 l3 mcters), including somc oPen-
 6 ADVANCED DAM ENGINEERING FOR DESIGN, CONSTRUCTION, AND
REHASILITATION
work boulder-gravel strata. Seepage control measures in_
cluded an inclined core, an extensive impenrious blanket
upstream from the dam, grout and drainage cunains, a
drainage blanket, a collector gallery, drainage wells, and
some grouting of foundation openwork. Undeneepage and
foundation behavior were monitored and euatuatea by
computerized methods.
lmproved Embankment Construction Control
A compaction meter has been developed to improve control
of embankment constrirction. In conjunction with elec-
tronic guidance and roller position detection, the meter en_
ables automatic recording of the degree of compaction at_
tained in the dam.
lnnovative Approaches in Rockfill Engineering
A principal feature of modern rockfill designs is {he com_
paction of the rockfill by a vibratory roller. Another sig_
nificant improvemenr is in the zoning of the fill. Importait
advances also have been made in the construction of con-
crete face slabs for rockfills.
Reinforced rockfill dams designed to withstand overtop_
ping have been constructed by use of steel surface mesh
and deep anchor bars.
In the Soviet Union, rockfill dams have been constructed
by directed blasting effected from mine chambers in the
canyon walls.
In the construction of the High Aswan Dam in Egypt, a
large volume of rockfill was dumped under water by means
of hopper barges and dump barges. Sand was placed in the
dam body by pumping a suspended mixrure through pipe_
lines. Sand in the foundation layer of the dam was com_
pacted under water by vibrating units operated from float_
ing installations.
Advances in Concrete Dam Technology
New data are available on the tensile strength of concrete.
Relationships between static and dynamic tensile strengths
have been defined.
Silica has proved valuable in improving the properties of
concrete. Because silica particles are two orders of mag-
nitude finer than those of common portland cement, they
provide a microfiller benefit as well as a pozzolanic effeci,
reducing the concrete's penneability and increasing its
strain capacity.
Use is made of a thermal-stress testing machine in which
a concrete specimen is cooled until it cracks. Concrete per_
formance can be studied with any temperature variation and
with different conditions of restraint.
Simplified procedures are used for analysis of the re_
sponses of concrcte gravity dams to scismic loading. in-
cluding the effects of interactions of the dam with the res-
ervoir and with the foundation. as well as rhe effect of
reservoir bottom absorption.
A method has been introduced for dynamic analysis of
arch dams, including hydrodynamic effects. This entails a
finite-element model in which the dam and fluid domain
are treated as substructures, and the effects
of foundation
ffexibility on hydrodynamic pressures are taken into ac-
count.
Linear elastic fracture mechanics is used in analysis of
crack formation and propagation in concrete dams under
dynamic loading.
Special concretes were used to solve severe problems
due to cavitation and erosion in concrete structures at Tar-
bela Dam. Large volumes of roller-compacted concrete
were placed to protect structures and rock slopes and foun-
dations. Fiber-reinforced concrete was used in the repair of
spillways and stilling basins.
Synthetic compound mixes and polymer impregnation
have been used in the repair of concrete damaged by cav-
itation or erosion.
Techniques are now available for repairing cracked con-
crete dams by grouting with synthetic resins. Very fine
cracks can be penetrated by these low-viscosity materials,
which, when ser, are deformable and highly adhesive.
Stressed steel anchors are used effectively in improvin_e
the stability of concrete dams. Anchorage is enhanced by
new techniques for protection against corrosion and for in-
creasing the bond.
Embedded, verrical, hydraulically inflatable flat jacks
were used to prestress a concrete arch dam (Nambe Falls
Dam, in New Mexico) to control tension at an acceptable
limit. Measurements of structural performance confirmed
that the prestressing was successful.
The technology of roller-compacted concrere dams has
advanced rapidly, as major structures of this type have been
constructed in several countries. progress has been made
in improving the impermeability and the durability of these
dams, while maintaining their economic advantage.
lmproved Techniques for Hydrologic and Hydraulic
Analysis
Mathematical models are used for flood forecasting and
reservoir operdtion simulation. Risk-based analysis has
gained some acceptance for selection of inflow design
floods.
Facilities f'or increasing flood discharge capability are
under continuing developnlcnt. Labyrinth weirs, for ex-
anrple, have been used to increase spillway capacity eco-
nomically and reliably. They are particularly suitable at ex-
isting danrs whcre the head ancl the space are lintited.

lncreased emphasis is placed on aeration of spillways
and outlets to prevent cavitation by high-velocity flow.
Refinements in lnstrumentation
Instrumentation and surveillance of dams are enhanced by
use of advanced vibrating wire instruments and electro-op-
tical distance-measuring equipment.
Significant improvement has been made in measurement
of total stresses in an embankment by using remote-reading
inclinometers attached to earth pressure cells to register any
changes in inclination after installation.
INTROOUCTION 7
lnnovations are being put to work in ncw projects as wcll
as in the evaluation and rchabilitation of older facilities.
Although most of the advanced concepts have been wc'll
proved. somc still must stand the test of time. As in most
situations wherc complex problems arc confr<rnted, new'
ideas in the engineering of dams do not always win im-
mediate acceptance. Opposing schools of thought may e\-
ist even aftcr thorough examination and trial by experi-
ence. Some problems must be recognized as amenable to
more than one solution. In any case. the exposure to cre-
ative thought fostered by timely professional interchan-ee
brings many benefits.
 LESSONS FROM NOTABLE EVENTS
LnunpNcp B. Jerurs
Chi ef En gi ne e ri n g Ge ol o gi st
Califurnia Department of Water Resources
(Retired)
RoseRr B. Jnusrru
Consulting Civil Engineer
Be llin gham, Washingt on
INTRODUCTION
Historical study of dams conceived in earlier times is es-
sential. To continue advancing, the engineering profession
must periodically review past problems and the lessons that
they taught. Candid sharing of information on failures as
THE BALDWIN HILLS RESERVOIR FAILURE
THoua,s M. Leps nNn RossR'r B. JnNsrN
On December 14, 1963, at about ll:15 e.pr., an unprec-
edented flow of water was heard in the spillway pipe at
Baldwin Hills Dam in Los Angeles, California. The water
came from drains under the reservoir lining.
At approximately l:00 p.n., muddy leakage was dis-
covered downstream from the east abutment of the dam,
which formed the north side of the reservoir. At 2:20 p.M.,
lowering of the reservoir water level revealed a 3-flt-wide
break in the reservoir's inner lining. A futile attempt was
made to plug the hole with sandbags. Water broke violently
through the downstream face of the dam. By 5:00 e.u.,
the reservoir had emptied, revealing a crack in the lining
8
GeoncE A. Krenscn
Geological Consultant
Kiersch Associates, Inc.
Tucson, Arizona
THoures M. LEps
President
Thomas M. Leps, Inc.
Consulting Civi I Engineering
Menlo Park, Califurnia
well as successes is needed. In fact, some of the most val-
uable learning has come from projects where errors have
been clear in retrospect. The following case histories are
representative of the body of knowledge that has been ac-
cumulated in the interest of the future safety of dams.
extending across the reservoir bottom in line with the breach
in the dam (Fig. 2-l).
Geologic Setting
The Baldwin Hills, on the southwest edge of the Los An-
geles Basin, are an expression of the Newport-lnglewood
upliit, a series of structural domcs and saddles extending
about 40 miles (64 km) bctween Beverly Hills and Newport
Beach. They are composed of sedintcntary formations,
principally of marine origin, overlying crystalline schist at
depths of 10,000 to 12,000 ft (3050 to 3660 meters).

Fig. 2- 1. Baldwin Hills Reservoir after failure.
The Inglewood Fault, the most noftherly of four princi-
pal faults of the uplift, is about 500 ft ( 152 meters) west of
the resewoir. The Ne*'port-Inglewood uplift remains
seismically active. However, at the Seismographic Labo-
ratory of the California Institute of Technology. l5 miles
(24 km) northeast of the reservoir, there was no report of
any earthquake considered large enough to cause inertial
damage at the projecr during the period 1950-63.
Several minor, steeply dipping faults were mapped in the
Baldwin Hills during construction. Three of these, desig-
nated Faults I, II, and V, pass through the reservoir, an-
gling away from the Inglewood Fault.
The reservoir foundation consisted of sediments that
were susceptible to densification and erosion. During con-
struction the formations were seen to be intensely jointed.
Most of the joints were tight, but a few had gaps of as much
as I in. (6 mm).
Design and Construction
Construction began on January 13, 1947, and was com-
pleted on April 18, 1951. Located at the heacl of a north-
ward-draining ravine, the reservoir was formed by thc dam
on the north side and compacted earth dikes on the othcr
sides. Designed as a homogeneous earthfill, the dam was
232 ft (71meters) high and 650 ft (198 meters) long.
The embankments were constructed of materials exca-
LESSONS FROM NOTABTE EVEI{TS 9
vated from the reserv'oir bowl. To provide a uniform floor
slope, part of the subgrade was compacted earth. The de-
sign incorporated underdrain systems and a reservoir lining
(Figs. 2-2 and 2-3).
Below the earth lining lay a 4-in. (102-mm) Iightly ce-
mented pea-gravel drain blanket, with a system of 4-in.
clay tile pipes placed beneath it to convey the leakage to a
drainage inspection chamber. The pea-gravel blanket tas
capped with a l-in. porous sand gunite layer to prevenr in-
filtration by soil particles. The lowest member of the lining
was an asphaltic membrane about j in. thick. sprayed on
either the natural formation or compacted fill. Cotton fabric
with an open weave served as reinforcement of the mem-
brane where needed, mostly on the slopes and in the ditches
for drain tile. The designers realized that the integrity of
the reservoir would depend upon the impermeability of the
asphaltic underseal.
Two separate underdrain systems were provided. one to
drain the foundation under the earth embankments and the
other to collcct seepage passing through the earth lining
and convey it through a central observation and measure-
ment chambcr to an outfall pipe.
In the bottom of thc ravine undcr the ntain embankment
was an open-jointcd l2-in. (305-nrm) clay tile foundation
drain, its uppcr half covercd with lightly ccmented pea
gravel to permit water entry. This drain passed through
successive manholes where seepage flow was observed.
1O AOVANCED DAM ENGINEERING FOR DESIGN, CONSTRUCTION. AND REHABILITATION

CN

:I =
lrj
i/ F
tt
,th
:r
::
.!
:.
i:
l-(, rB

I
o3! !i
gi' ,il
3r: a
":i E
o
rl
il

f:' -
lt o
I
-v \.

it.'"'\

tr

I'l

ttI
\l a4

.z
./ :, o
I
\-_ N
d

U l!

t./
/
\.42^.-
,/i.i
t l-/-
,-
-_-,;iH
f, ./
!t //
It/ /'o''..---

EL.t39
{. I
/lsPH Llrc PAVtt{c
€L.tr.57
SPHAL?IC
4" CLAY ?ILE ORAIN
llr co{cFEtE cRAoLE
( RESERVOIR EOTTOM Ot{LY
4' c€u€llrEo PEA-GRAVEL
lla'lsexrurrc uEuBR^xE
LYPICAL SECTION THROUGH LINING
Fig. 2-3. Baldwin Hills Reservoir lining.
Horizontal and vertical holes were drilled into the dam
foundation and were sand-filled. They were connected to
4-in. tile drainpipes leading to the l2-in. bottom pipe,
which discharged into the city storm drain immediately
downstream from the dam. Radiating from each of the
manholes were other horizontal drill holes, backfilled with
sand, to intercept seepage from the foundation.
Baldwin Hills Reservoir had a storage capacity of 897
acre-feet (l.l million m3) and was served by mortar-lined
steel inlet and outlet conduits in tunnels through the east
side. A 57-in. (145-cm) inlet line extended through the gate
tower, supplying two circulator pipes leading to the west
side of the reservoir. The tower had gates at various levels
for discharge into the 66-in. (168-cm) outlet line.
Surveillance
The owner, the Los Angeles Department of Vy'ater and
Power, conducted a comprehensive program to measure
performance of the project. The reservoir was in service
continuously from July 195 I until failure on December 14,
1963, except for a short time in 1957 when it was drained
for cleaning and repairs.
A caretaker was on duty regularly at the reservoir from
7:45 e,.u. to 4: 15 p.M., seven days a week.
The Foundations and Structures Maintenance Section
devoted one full day each month to inspection of the res-
ervoir.
Leakage was regularly measured and sampled by hy-
drographers of the Water Operating Division, whose activ-
LESSONSFROM NOTABLE EVENTS 11
lsPHlLttc PAVIio
KE
M € M BRAN€
PEA.GRAVEL ORAIrl
SECTION ON RESERVOIR SLOPE
)
oRAt]a
RESERVOIR LINING
ities during a typical weekly inspection included entering
Manholes A and B on the l2-in. tile drain and measuring
flows from the horizontal drain holes. They also examined
the spillway catch basins and the storm drains.
Flows of the reservoir underdrains were measured at
weirs in the inspection chamber before discharge into a 24-
in. (61-cm) outfall line, which connected to the spillway
pipe.
Flows in the embankment foundation drains were mea-
sured monthly. Throughout the project operation no flows
were observed in the drains for the embankments on the
east, west, and south sides.
Periodic inspections were made of observation wells at
the reservoir perimeter. Reportedly there was never any
water in these wells.
History of Operation
During the first reservoir filling, begun on April 18, 1951,
discharge from the underdrains increased substantially.
Reservoir inflow was discontinued on May 2 with the stor-
age level about 22 ft (6.7 meters) below the spillway. Re-
pairs were made. Reservoir filling was resumed on June
r8,1951.
In the early years of operation following the initial re-
medial work in 1951 , the underdrains required much main-
tenance. Appreciable volumes of asphalt, obviously from
the asphalt membrane, flowed through the system from the
west underside of the reservoir.
Calcium carbonate deposits developed in the drains, re-
 12 ADVANCED DAM ENGINEERING FOR DESIGN. CONSTRUCTION. AND REHABILITATION
quiring frequenr cleaning. Clogging, and possibly dis-
placement of the drain tile, caused a reduction in rhe toral
seepage entering the inspection chamber. Seepage varied
over the years from slightly more rhan 23 gpm (g7 li-
ters/min.) to a low of approximately 7 gpm (26 li-
ten/min.), and measured about 8 gpm (30liters/min.) in
early 1963. Wide fluctuations of flow were observed in the
individual underdrains, with much of the variation un-
doubtedly attributable to repeated acid treatment of the
drains.
On October 29, 1951, a crack was discovered in the
drainage inspection chamber near Fault I. Strain-gage
points were set, and regular n'leasurements of the opening
were begun. The crack opened enough that the steel rein-
forcing bars in the wall were visible. Leakage through the
opening was reportedly negligible prior to the reservoir
failure.
The five drains entering the inspection chamber from the
west were encased in concrete for a distance of about 20 ft
(6 m). At rhat location they were below the
|-in. 1O-mm;
asphalt membrane. On September 5, 1952, the concrete
encasement was found to be cracked approximately 12 ft
(3.7 meters) from the discharge end of the west toe drain.
From the time that the crack opened, it was reasonable to
assume that there was leakage into the foundation at that
point.
In the period March 13-16, 1957, the reservoir was
emptied, and the lining cleaned and checked. Inspectors
found some cracking in the thin (about {-in.y cement coat-
ing on the asphaltic pavement, presumably due to creeping
of the paving on the reservoir slopes. Although no sub-
stantial separation was found in the pavement, in some lo-
cations there was an overthrust of as much as 2 in. (5 cm).
The reservoir bottom was found to be in generally sound
condition.
In the weeks immediately preceding failure an apparent
uplift developed in the inlet tunnel, the gare tower, and the
part of the inspection chamber east of Fault L This reversal
of the settlement trend was similar to others that had oc-
curred occasionally during the life of the reservoir, but it
was larger than any previously recorded.
In the final year of operation, the flows from discharging
horizontal drains under the main dam varied rapidly, de-
clining to zero and then in some cases increasing to the
earlier rate, followed by continued fluctuations.
Postfailure Conditions
After the failure, about 2 in. (5 cm) ol fine silt and clay
covered the reservoir floor.t There was a continuous crack
approximately parallel to and near the toe of the east slope.
Vertical displacement averaged about 2 in. (5 cm), but was
as much as 7 in. (18 cm), with the west side of the crack
lower than the east side. The crack extended up the south
slope of the reservoir but with very little offset. Several
sinkholes could be seen in the reservoir bottom along the
continuous crack.
of
Inspection preexisting cracks in the inlet tunnel
showed no apparent chan_9e in the conditions of any of
them.
On December 18, 1963, the drying of sedimenr on the
floor disclosed a north-south-trending crack near the res-
ervoir center. Subsurface investigation showed this crack
to be at Fault V. There u,as very little displacement, but
water apparently had been passing through the crack into
the earth lining.
Postfailure investigators explored Faults I and V inten-
sively. Thirteen test pits, two shafts, and 256 ft (78 meters)
of tunnel were excavated. Also, drilling was done at the
site to obtain samples of materials, for tests performed on
both undisturbed and disturbed specimens of the natural
and compacted soils.
Excavations at Faults I and V revealed downward dis-
placement on the west side, which was evident in the res-
ervoir paving but more pronounced at the pea-gravel drain,
where offsets of as much as 7 in. (18 cm) were observed.
These were generally extensions of rupture in the founda-
tion.
At Fault I, cavities were discovered beneath the pea-
gravel drain in some of the excavations.
At Fault V. in the north face of an exploratory trench, a
hole found about 8 ft (2.4 meters) below the pea-gravel
drain opened into a cavity about I I ft (3.4 meters) long, 3
ft (1 meter) high, and as much as 2 ft (0.6 meter) wide. It
extended north along the west side of Fault V. The pea-
gravel drain at the fault was displaced, but it had not col-
lapsed. A slight calcification on the surfaces of the large
cavity suggested that it had developed over an appreciable
time.
Excavation at Fault I disclosed cavities in the natural
formation as deep as 47 ft (14 meters) below the pea-gravel
drain. These cavities apparently had existed for an ex-
tended period.
In a drift excavated along Fault I, investigators encoun-
tered numerous cavities, some with maximum dimensions
of several feet. The relatively smooth and straight fault
planes suggested that separation of foundation blocks was
the primary action in creating the gaps. Fine sand was flound
in some of the openings between the fault planes.
Some flra,ements of the asphaltic membrane under the ce-
mented pea gravel contained small holes that could have
allowed passage of water.
Settlement records indicatcd that a local trough of max-
imum settlement crossed the rcservoir parallel to and just
west of the trace of Fault V.
Surveys of the inlet and outlet tunnels, gate tower, and
inspection chamber were conducted in the week after fail-
ure. With respect to measurements made on November 20,

1963. there was a relative uplift of 0.01 ft (3 mm) at the
east portal,0.ll ft (34 mm) at the tower, and 0.17 ft (52
mm) in the inspection chamber east of Fault l. There was
no measured change in the chamber elevation west of Fault
I. The top of the gate tower had moved to the east 0. 12 ft
(37 mm). Movement of the gate tower base had been gen-
erally downward over the life of the structure. reachin-e a
maximum of 0.55 ft (168 mm) before a sudden uplift of
0.I I ft (34 mm) at about the time of failure.
The lnglewood Oil Field
The Inglewood Oil Field, discovered in September 1924.
lies under the western half of the Baldwin Hills area. It
covers about 1200 acres and in 1963 had more than 600
producing wells. The field adjoins the reservoir site on the
south and west, the nearest reported production at the time
of reservoir failure being from three wells within 700 ft
(210 meters) of the south rim.
The geological structure of the field, the nature of the
oil-bearing deposits. the relatively shallow depth, and the
solution gas drive that prevailed in the early stages of pro-
duction provided an environment favorable for subsidence.
Production at the Inglewood Oil Field commenced in
1924, and repressurization of the field was started in 1954.
Ground Movement
In 1917, the Depanment of Water and Power established
benchmarks at the site of Centinela Reservoir. a proposed
storage facility about 3000 ft (900 meters) southwest of the
Baldwin Hills damsite. In 1943 levels again were run to
some of these points, and the changes since l9l7 deter-
mined. This and subsequent surveys indicated a general
regimen of subsidence in the area (Figs. 24 and 2-5). In
his extensive investigations. Leps: estimated the year of
initial subsidence as about 1924.
The poini of maximum cumulative subsidence in 1963
was about 3000 ft (900 meters) northwest of the intersec-
tion of La Cienega Blvd. and Stocker Street, over the most
productive part of the Inglewood Oil Field. The total area
of subsidence resembles an elliptical bowl with its center
about 0.5 mile (0.8 km) west of the reservoir and its eastern
periphery extending beyond La Brea Avenue. Subsidence
at the reservoir site aggregated about 3 ft (0.9 meter) within
the 1917-63 period. Its southwest cornerhad dropped more
than the northeast corner. Between 1947 and 1962 the el-
evation difference between these rcscrvoir corners in-
creased by about 0.5 foot (15 cm).
Triangulation surveys in 1934. 1961, and 1963 sho*'ed
stations in the Baldwin Hills to be moving laterally in thc
general direction of the subsidence elepression. There was
a progressive 0.4-ft (12-cm) elongation of the northeast-
southwest diagonal of the reservoir between 1950 and 1963.
TESSONS FROM NOTAELE EVENTS I3
ln an arca bctwecn 0.5 mile (0.8 km) and I mile (1.6
km) southeast of the reservoir. several open cracks were
detected in the grcund as early as May 1957. They mea-
sured up to 1500 ft (760 meters) in length. and were similar
to the 1963 reservoir cracks in terms of vertical offset.
opening. orientation, and lack of longitudinal displace-
ment. [n general. both the reservoir cracks and those to the
southeast panlleled old faults and joints. They developed
in areas where tension could be expccted as a result of
warping of the ground surface.
Analysis of Failure
Several investigators have offered premises about the cause
and mechanism of the reservoir failure. Althou-eh they share
areas of substantial agreement, some of their differences
have not been fully reconciled.
Hamilton and Meehanr concluded in l97l that fluid in-
jection caused shear displacements alon-q Fault I. and that
rupture propagating to the surface sheared the earth lining.
Casagrande et al.a were not convinced that any signifi-
cant fault movement had taken place under the reservoir
during its life. They believed that differential settlement
occurred and could be explained by the greater compress-
ibility of fractured and loosened material immediately west
of each fault. They assumed that during pre-Holocene tec-
tonic activity on Faults I and V the do,*'nthrown (west)
block was dragged along the fault plane so that the mate-
rials near the surface on the downthrown side were loos-
ened. Because the Inglewood formation is particularly fri-
able, they regarded these effects as severe. They therefore
concluded "that there was sharp differential compression
of the foundation soils across Faults I and V. that this dif-
ferential compression was initiated already durin-e the first
partial fillin-e of the reservoir in 195 l. and that it increased
gradually during the life of the reservoir."
Lepsl has stated that "the loosely articulated nature of
the fault blocks has represented a foundation environment
under the reservoir site which has been extremely sensi-
tive, and responsive, to the localized but substantial
changes in subsurface stresses caused by a subsurface, salt
water injection program begun on a pilot basis by the oil
field operators in 1954. and intensively pursued from
1957...."
The papers by Lepss and Hanrilton and Meehanl dem-
onstrated the apparent influence ofoil field repressurization
on the subsidcnce trend at the reservoir. The surveys of
facilities on the East Sub-block showed clearly the slowing
and short-term reversals of the downtrend. corrclating with
fluid injection. The cited papers concludcd that differential
vertical shearing at Faults I and tt was accentuated by the
reprcssuring. Recorded rebouud east and south of the res-
ervoir, and possibly some of the ground cracks in the en-
virons, appear to support this argument.
14 ADVANCED DAM ENGINEERING FOR DESIGN, CONSTRUCTION, AND REHABILITATION

15
!i IIii
::3t
?fr
!i tt!: EEq ,r
,i,;.1 ,-gi
,iillii*iiiiii .rl
;t!
iJl
!rrt:;
,iil;iii;
I. n[
9zz. r-l
tiB= !il
I i..i t
I
i'
i: r:!ti
ir i i'l EHd'I
lJo3 !l
ga

o
o

\2 N

d
u
E
 TESSONS FROM NOTAELE EVENTS I5

YEAR

F
E
UJ
TWO POSSIBLE PATHS OF
PBM 3I (BALOWIN AUX.)
(,
z SUESIOENCE OF PBM I22
6
F
u,
tJ
lr 68 (BM'OO')
z_

lr,
L'
z
LJ
I
al,
o
,
ah

trj
2
F
J
, O OATES WHEN ELEVATIONS WERE TAKEN
=
3
o NOTE, DASHEO LINES INDICATE POSSIELE PATHS OF
SUBSIOENCE WHERE NO OATA ARE AVAILABLE

Frofr Ilt LGp3. 1964

SUBSIDENCE OF BENCH MARKS
BALDWIN HILLS AREA
Fig. 2-5. Subsidence of benchmarks. Baldwin Hills area.

During drilling of oil wells south of the reservoir, cir-
culation of drilling fluids reportedly was lost in the upper
several hundred feet. Assuming this to be indicative of
conditions under the reservoir, ready channels for seepage
to great depth may have existed in the foundation, partic-
ularly at the faults.
There is evidence that gaps had developed between
foundation blocks over a long period before the reservoir
failure. They were seen and photographed at Faults I and
II in an excavation near the gate tower site in 1948.
Further separation of the fault planes is believed to have
occurred during the years of operation, which would tend
to cause failure of the brittle underdrains, rupture of the
asphalt membrane, and consequent wetting of any poorly
consolidated materials in the fault openings.
In general, postfailure measurements showed the fault
planes to be separated on the order of 0.25 in. (6 mm) to
0.5 in. (13 mm). In some places, the openings had been
enlarged by erosion, which could be attributed to rainwater
infiltration in years prior to construction, reservoir seepa-qe
in the period l95l-63, or outrush of water during the fail-
ure.
Separation of the fault blocks due to tension from warp-
ing of the subsidence bowl would not necessarily have re-
sulted in differentially compressible materials of the na-
ture. and to the degree. envisioned by the tectonic drag
theory. However, overhanging fault-block corners that ex-
isted near the ground surface at the time of construction
might have been broken by mobile equipment so that foun-
dation material was forced into the accumulated fault gaps.
Also. the wetting of the foundation could have made the
corners of the overhanging foundation blocks susceptible
to further breaking under the reservoir loading. All of this
would have been conducive to localized settlement.
Heavy construction equipmcnt also may have caused ini-
tial defects in the reservoir lining. Construction photo-
graphs show a scraper, a truck. and a motor grader oper-
ating directly on the cemented pea-gravel underdrain. A
 16 ADVANCED DAM ENGINEERING FOR DESIGN, CONSTRUCTION, AND REHABILITATION
tracked crane was run over the asphaltic membrane, on ply-
wood strips, during placement of the pea gravel. Much of
this was membrane without fabric reinforcement.
Until the final episode of failure, the relatively small
amounts of water that probably were entering the founda-
tion could have been expected to pass through the perme-
able sedimentary materials and decply into the fault gaps
without developing sustained pressures of a magnitude that
would move foundation blocks. On the other hand, the out-
rushing water at time of failure would have introduced
much higher pressures. Although they might not have been
able to move massive blocks, they might have rcduced in-
terplanar restraint enough to allow a small rebound, and
they could have caused lateral shifting of the relatively thin
foundation units between closely spaced faults such as that
between Faults I and II. The Inglewood formation is in-
tensely jointed, so that there might not be substantial resis-
tance to such horizontal movement.
In summary, the reservoir and its immediate environs
were subjected to many adverse forces, including horizon-
tal and vertical displacement due to subsidence; local
breaking of the weak foundation; some erosion at the faults;
and rebound effects due to oil field repressurization, res-
ervoir loading and unloading in l95l and 1957, and the
final inrush of water into the Fault I-II zone at time of
failure.
With the benefit of the data that the failure provided, of
course, the foundation problems would be predictable in
such a setting. Under the same circumstances now, the set-
tlement and seepage pattems that were observed at Baldwin
Hills would signal the need for decisive corrective action.
From a practical standpoint, present methods of interpret-
ing perflormance data are not substantially better than those
available in 1963. The instrumentation and the measure-
ment procedures employed by the Department of Water and
Power were adequate. In retrospect, they passed the test of
showing what was happening.
Facing the same site conditions now, and armed with the
knowledge of what went wrong at Baldwin Hills, designers
could make several improvements. The obvious first step
would be to avoid rigid drains so close to the water face
and to the unstable and erodible foundation. A paved earth
lining similar to that in the original design would be ac-
ceptable, but preferably of soils with higher plasticity. An
impermeable, reinforced, multi-ply synthetic membrane,
placed on a well-prepared subgrade, might provide an ef-
fective underseal. Any drain layer over this would have to
be flexible and filter-protected. Overexcavation at the faults
and backfilling with compacted clay would provide addi-
tional protection from cracking. A sandwiched lining of
this type would be capable of appreciable deformation
without rupture. However, it probably could not be guar-
anteed to remain completely watertight in the long term
while its foundation continued to be strctched as a conse-
quence of ar€al subsidence. Aging of the synthetic mem-
brane might be a problem over a period of many years. The
designers of any facility at the sire also would have to be
mindful of the seismic porenrial of the area.
Whether the engineers who conceived the original plan
should have anticipated the nature ofthe foundation breaks
is not easy to judge fairly in rrtrospect. The project was
well reported in the technical journals and had few critics
at the time. There was little precedent to offer guidance in
the 1940s. The Baldwin Hills case, considering all its ele-
ments, is still unique in the history of dams.
Lessons Learned
l. Foundations in erodible rocks must be thoroughly
explored to disclose any preexisting cavities or other
defects.
2. The total prevention of leakage into a reservoir
foundation over the lifetime of the facility may be
unattainable under usual circumstances.
3. Associated faults that lie in close proximity and sub-
parallel to an active fault should be reganded as sus-
ceptible to movement in a seismic event.
4. The possibility of differential fault movement un-
related to tectonic activity must be considered.
5. Potential effects of ground subsidence must be rec-
ognized in designing dams and reservoirs.
6. External causes and effects of subsidence must be
closely monitored.
7. Foundation discontinuities should be given special
treatment in construction.
8. Rigid buried elements such as cemented drains
should not be incorporated into designs where dif-
ferential settlement is a possibility.
9. Drains should be amply sized and provided with ac-
cess, where possible, to facilitate maintenance.
10. Application of sprayed asphalt as a reservoir seal
must be questioned as to its long-term effectiveness.
ll. Earth linings preferably should have appreciable
plasticity.
12. Erodible embankment and foundation elements must
have adequate filter protection.
13. Structures placed across faults should be conserva-
tively designed to accommodate predictable move-
ments.
14. The use of heavy construction equipment must be
carefully controlled to avoid damage of critical res-
ervoir features on soft foundations.
15. Surveillance of a rcservoir must be extended to its
environs and to the consequences of adjacent de-
velopments and physical changes.

THE FAILURE OF MALPASSET DAM
LeunrNce B. J,tltrs
lntroduction
Malpasset Dam failed suddenly during the night of Decem-
ber 2, 1959, althou-gh a few hours earlier engineers had
detected nothing unusual at the site. The collapse occurred
in a burst, which destroyed practically the entire dam. re-
sulting in the loss of several hundred lives and many mil-
lions of dollars in property damage. Subsequent trials in
French courts-lasting six years and involving extensive
investigations by experts-concluded that the failure was
unpredictable.
As with most catastrophic dam failures, much evidence
was swept away by the floodwaters. Consequently. hy-
potheses for the causes and mechanisms of failure are nec-
essarily based in part on conjecture. This review of plau-
sible explanations accounts for all of the factual
observations assembled during the ensuing years' consist-
ent with current knowledge and concepts of applied geo-
logical sciences and rock engineering. In addition, the re-
sults of many years of intensive investigation. conducted
largely by the designers of the dam. are summarized,l as
are the findings of more recent studies, reported in August
1985.2
The writer spent three days on the site in May 1964 re-
viewing the site conditions, causes, and evidence of failure
*:
qi
Fig. 2-6. Malpasset Dam upon its complction in 195.1. Courtesy of Coyne et Bellier. Burcau D'lngenieus Conscils' Paris
LESSONS FROM NOTASLE EVENTS 17
with Professor George Kiersch and mcmberu of thc Coyne
et Bellier staff. including Picrre Londe.
Description of the Dam
Malpasset Dam (Fig. 2-6) was located in a narrow gorge
of the Reyran River approximately 7.5 miles (12 km) up-
stream of Frejus, a resort community on the shores of the
French Rivieru. The dam was a double curvature arch of
218 ft (66.5 meters) maximum height, with crest length of
728 ft (222 meters) and thickness that varied from 4.9 ft
(1.50 meters) at the crest to 22.2 ft (6.77 meters) at the
base of the central block. A thrust block on the left abut-
ment measuredT2 ft (22 meters) in length and 21.3 ft (6.5
meters) in width. and rose 36 ft ( I I meters) above the foun-
dation concrete. A concrete win-e wall protected the thrust
block from the full force of the reservoir. Other features
included an ungated spillway 95 ft (29 meters) long at el-
evation 329.3 ft (100.40 meters) located on the crest of the
arch. a bottom outlet 4.9 ft (1.50 meters) in diameter con-
trolled by a butterfly valve, and a reinforced concrete apron
to protect against scour at the downstream toe of the arch.
A water intake on the left bank at elevation 261 ft (79.5
meters) passed through the dam, emptying into a float
..,......|
t
 18 ADVANCED DAM ENGINEERTNG FOR DESIGN, CONSTRUCTION, AND REHAEILTTATION
chamber, called a "danaide," that regulated the waterdis-
charge.
The rcservoir capacity at the spillway crcsr was 41,300
acre-feet (50.9 million m3).
Water was stored for irrigation, domestic, and municipal
use.
Geology3
The dam site is located in the Tanneron Massif (Fig. 2-7),
the right abutment consisting of massive gneiss while the
left abutment rocks are more essentially fresh schistose in
structure and variously altered in composition. Postfailure
microscopic examination of rock specimens disclosed lat-
tices of microfractures open to several microns with an av-
erage mesh of several millimeters. On-site inspection re-
vealed two types of joints, a denser mesh of 0.8 to 1.2 in.
(2 to 3 cm) and a second set of 8 to 20 in. (20 to 50 cm).
Shears and faults are numerous in the dam foundation, and
are oriented in the same directions as the jointing. Water
l,u __.
E(
-'-r1l.rlr
MASSIVE AUGEN GNEISS WITH
LOCALSCHISTOSE ZONES
/ A horizontal displacement of 0.6 in. (1.5 cm) of the base
JJl5o"'
,\ LEFT BANK
\tl F"'
NS TYPE FAULT
\ \
Fig.2-7. Malpasset dam sirc lies wirhin thc crystallinc Tanneron massive.
At thc site the rock is generrlly gneissic, with a tendency toward schistose
structure in the left bank and lower pan of the right bank. Shown by
hachures is thc seat of the n:ck weilge thar supponcd the left end of the
dam, which was dislodged. shattercd, and swept away during the failure
of the drm. Modified from Bellierr and Londe.r
pressure test results averaged less than one Lugeon.r Con-
sequently, the foundation was not grouted. except in the
contact zone immediately below the concrete block.
East (left) of the river channel the dam rcsted on a wedge
of gneiss. Prior to the failure it was not known that this
wedge was detached from the underlying rock mass by two
converging discontinuities: an upstream surface consisting
of shear planes oriented roughly parallel to the schistosity,
and a downstream surface consisting of a clay-coated fautt
plane rvhose existence was not suspected until it was dis-
closed by erosion during the flood following the rupture of
the dam. Although the wedge of rock was broken and swept
away by the force of the flood. the seat upon which it resred
remains (Figs. 2-8 and 2-9).
Preconstruction geological investigation included in-
spection of the site by a professor of geolo_ey,a eight grout
holes 65 to 100 ft (20 to 30 meters) in depth, and exami-
nation of the foundation excavation. No core drilling rvas
performed. The fint geological investigation was based on
the supposition that a gravity dam would be constructed. It
is not clear whether the geologist was duly informed u'hen
the design was changed to an arch dam.5
---
Prefailure Observations
Construction of the dam commenced in the spring of 1952
and was completed in April 1954. Reservoir filling began
in late 1954 but was delayed because of low rainfall. Ex-
ceptional storms in November 1959 filled rhe last l3 ft (4
meters) of storage in three days. During the filling, mea-
surements were made of the arch behavior on four occa-
sions, in September 1955, July 1956, July 1958, and July
1959. The last observation was made four months before
the collapse of the dam. All measurements were obtained
by geodetic sun'eys on 28 targets fixed on the downstream
face of the dam. No instruments were installed in the dam
or its foundation to detect or record deformations.
of the dam was indicated in the last measurement before
the failure. A comparison between displacements measured
in 1959 with corresponding computed displacements based
on a ratio of E,/E, = l0 (where E" : modulus of con-
crete, and E : modulus of rock) showed that the rock
actually yielded much more than the computed values.
Vertical displacements along the base and crest of the dam
indicated a general tilting of the arch, the right side being
raised slightly more than the left. The uplift was mosr pro-
*A criterion for relating the resuhs of watcr prcssure tests in foundations
for concretc dams over 100 ft (30 nrctcrs) high to rhe necessity for grour-
ing. A unit Lugeon is equal ro a watcr loss of one liter/meter of hole/min.
(0.08 t gal. /ft of hole/min. ) wherc the apptied wlter prcssure is held at
l0 atm. According to this criterion. grouring is not necessary if *.ater
losses are less than one Lugeon.
 LESSONS FROM NOTAELE EVENTS 19

Fi,e. 2-8. postfailure view of the left abutment of the dam. Upper center-the thrust block and wingwall remnant. Right center-the danaide. The detached
.oncrete block (center) broke loose during the final stage of the collapse and toppled to its present posilion. It now rcsts on the surface of the fault along
which the rock wedge slid at the onset of rhe failure. The discontinuity that formed the upstream seat for the wedge (lower leti quaner; was created u'hen
the rock ruptured along planes of schistosity. The wedge was srnashed and swept away by the force of the escaping water. Lo*.er right-the remains of
the ancient Frejus Aqueduct constructed by the Romans. Counesy ofCoyne et Bellier. Bureau D'lngenieurs Conseils-

nounced, 1.6 in. (4.1 cm), at joint F. (Figure 2-10 shows
joint locations.) In retrospect, these discrepancies may be
interpreted as harbingers of problems within the foundation
of the dam. Some cracks were observed in the concrete
apron at the base of the dam a few weeks before the failure
but had not been reported. A week prior to failure, a small
stream of clear water appeared io the right wall of the gorge
some 65 ft (20 meters) downstream of the dam at elevation
262 ft (80 meters). This leak was later evaluated to be of
no significance in the failure.
In the afternoon of the day of failure, a group of advisory
engineers observed the foot of the dam and decided to open
the bottom outlet gate to control the rising reservoir pool.
No vibrations were noticed. At about 1800 hours, the care-
taker, who had been performing routine painting. left the
crest of the dam. Neither he nor the engineers had observed
anything suspicious. At 2ll3 hours, Electricite de France
registered a power outage on the Malpasset l0 kV line. The
first failure of a modcm concrete arch dam had occurred,
catastrophically.
The witnesses closest to the scene were 0.93 mile (1.5
km) downstream, their vision impaired by darkness. They
reported first feeling a trembling of the ground, followed
by a loud, brief rumble, then a strong blast of air. Finally.
the water arrived in two pulses, a wave that overflowed the
stream banks and then a wall of water, which the witnesses
barely escaped.
Postf ailure Observations
Only remnants of the dam remained (Fig. 2-ll). A re-
survey indicated that the dam remnants had shifted down-
stream, and that the magnitude of the displacement of each
part was proportional to its distance measured from the right
end of the dam. Also, a crevice had opened along the up-
stream face of the concrete arch at its contact with the right
abutment. It was apparent fmm these observations that the
dam had rotated as a solid body about an axis located in
the right abutment (Fig. 2-12). An exploratory shaft at the
foot of the right bank revealed a horizontal crack that ex-
tended bencath the dam close to thc rock-concrete contact.
(Fig. 2-13). The presence of this opcning suggested that
the <Jam had becn lified off its foundation, thus explaining
how it was able to rotate as an intact unit. Further evidence
ol this upliftingwas observetl at the danaide. which had
straddled the fault antl consequently was severed by the
 20 ADVANCED OAM ENGINEERING FOR DESIGN, CONSTRUCTION. AND REHAB]LITATION

Fi,e. 2-9. The left abutment viewed from upslream. The concrete thrust
block (extreme upper right) was isolated from direct contact wirh the res-
ervoir by a wing wall. a remnant of which still stands next to the thrust
Fig. 2-l l. Remnants of rhe concrete arch. Courtesy of Coyne et Bellier,
block. The seat of the rock wedge that supponed the left end of the dam
Bureau D'lngenieurs Conseils. Paris.
consisted. on the upstream side. of a discontinuity made up of cleavage
planes alon,g the rock schistosity. These smooth surfaces are shown in rhe
photo extending from ttre lower left to upper right quadrants. They are
aligned approximately parallel to the direction ofthe thrust ofthe concrere
arch. Courtesy of Coyne et Bellier Bureau D'lngenieurs Conseils, paris.

RIGHT BANK LEFT BANK

(tt) (ml A B
'100

300 9A

80

10

200 0-a_ EASE OF EXCAVATION

9S_

4q
BOTTOM OF LARGE WEOGE OF ROCK
1 00-
Fig. 2-10. A postfailure profile showing thc outline of the remains of the danr, looking upstreanr. The alphabetieal designations of thc arch joints and
intervening concrere blocks are shown along the crcst of ihe dam. Fmrn Bellier.t