See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/342317652

Dam Safety and Earthquakes

Article · January 2020

CITATIONS READS

8 9,364

5 authors, including:

Nasrat Adamo Nadhir Al-Ansari

Private University Consortium Ltd Luleå University of Technology
115 PUBLICATIONS 894 CITATIONS 1,140 PUBLICATIONS 16,342 CITATIONS

SEE PROFILE SEE PROFILE

Varoujan K. Sissakian Jan Laue

Komar University of Science and Technology Luleå University of Technology
252 PUBLICATIONS 2,271 CITATIONS 267 PUBLICATIONS 2,384 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Nadhir Al-Ansari on 19 June 2020.

The user has requested enhancement of the downloaded file.

Journal of Earth Sciences and Geotechnical Engineering, Vol. 10, No. 6, 2020, 79-132
ISSN: 1792-9040 (print version), 1792-9660 (online)
Scientific Press International Limited

Dam Safety and Earthquakes

Nasrat Adamo1, Nadhir Al-Ansari2, Varoujan Sissakian3, Jan Laue4
and Sven Knutsson5

Abstract

Earthquakes may cause failure or profound damage for dams. Factors contributing
to this are, magnitude on the Richter scale, peak horizontal and vertical
accelerations, time duration, in addition to the epicentral distance, nature of
foundation rock, criteria of the design, and finally, if appropriate type of dam and
materials has been used. Extensive lists of dam failures and damaged once are
presented with many case histories. Most failed dams were tailing dams or hydraulic
fill dams or small earth fill dams, which reflect the weight of the design and
construction factors. Embankment dams, normally, are less tolerant to ground
shacking than concrete dams. While rockfill and RCC dams have shown good
performance. The developments of design methods and criteria are traced here, from
the early use of the pseudoptotic method to the more rational dynamic analysis,
which is used nowadays making construction of very large safe dams in seismic
regions possible. The method adopts peak ground accelerations from anticipated
earthquakes as inputs to the analysis which produce a full spectrum of the factor of
safety during any considered event. This has led to increased use of seismic
instrumentation to produce seismographs of actual events in the free field, and on
dams hit by earthquakes for comparison with outputs of this analysis and for future
use for similar dams in similar circumstances, and to decide on rehabilitation
measures. The safety levels to which any dam is to be designed are defined in terms
of the Maximum Credible Earthquake, Safety Evaluation Earthquake, Maximum
Design Earthquake and other similar terms. Dam repairs after sustaining earthquake
damages are described in real cases and upgrading of older dams to withstand higher
expected seismic events are also treated here and supported by case histories.

1
Consultant Dam Engineer, Sweden.
2
Lulea University of Technology, Lulea, Sweden.
3
Lecturer, University of Kurdistan Hewler, and Private Consultant Geologist, Erbil.
4
Lulea University of Technology, Lulea, Sweden.
5
Lulea University of Technology, Lulea, Sweden.

Article Info: Received: June 9, 2020. Revised: June 18, 2020.

Published online: July 10, 2020.
80 Nasrat Adamo et al.

Keywords: Earthquakes, Seismicity, Epicentral Distance, Focal Distance, Peak

Horizontal Acceleration, Pseudo-static Analysis, Dynamic Analysis, Maximum
Credible Earthquake, Safety Evaluation Earthquake, Maximum Design Earthquake.

1. Earthquakes in General
An earthquake is the shaking of the Earth surface resulting from a sudden release of
energy in its crust and its uppermost mantle, which constitute its hard and rigid outer
layer, and, therefore, creating elastic waves known as seismic waves. Earthquakes
can range in size from those that are so weak that they cannot be felt to that violent
enough to toss people around and destroy whole cities. The seismicity, or seismic
activity, of an area are the frequency, type, and size of earthquakes experienced over
a period of time. Earthquakes and the release of its energy are caused mostly by
rupture of geological faults but also by other events such as volcanic activity;
landslides, mine blast, and nuclear tests. An earthquake's point of initial rupture
deep, down is called its focus or hypocenter, and the epicenter is the point at the
ground level directly above the hypocenter. The mechanism of this rupture is
attributed to the fact that the upper mantle of the Earth consist of seven or eight
major tectonic plates, (depending on how they are defined), and many minor plates,
which are in constant motion Figure 1 [1]. Where the plates meet, their relative
motion determines the type of boundary: convergent, divergent and transform.
Earthquakes, volcanic activity, mountain-building, and oceanic trench formation
occur along these plate boundaries (or faults) which form the seismic belts of Earth.
The two major seismic belts are the Circum-Pacific Belt, which surrounds the
Pacific Ocean, and the Alpine Belt, which stretches from the Azores through the
Mediterranean and Middle East to the Himalayas and Indonesia, where it joins the
Circum- Pacific Belt. A purely oceanic seismic belt lies along the mid-Atlantic ridge,
[2]. Along these belts the majority of earthquakes epicenters cluster, refer Figure 2
[3].

Figure 1: The Earth’s tectonic plates. (Source: USGS) [1].

Dam safety and Earthquakes 81

Figure 2: Distribution of Major Earthquakes in the World [3].

The location of any point on the earth surface close to any of these belts indicates
high probability of earthquakes occurrence at that point. It is the magnitude of any
earthquake, which represents the governing factor for the extent of damage that can
happen to structures, whether being dams, buildings, bridges or any other
infrastructure or lifeline. Local faults, even not related to these major seismic belts,
can also give rise to earthquake and result in such damages and even loss of life.
The obvious conclusion that may be drawn is that, it is the duty of the designing
engineer to ensure the safety of these structures against the impacts of earthquakes
by using building codes. In major structures such as dams, Atomic Energy Power
Stations and the like, a complete Seismic Study is not only justified but strongly
recommended.

2. History of Dams Failures and Incidents due to Earthquakes

During late history, not many dams have collapsed or very badly damaged
worldwide by earthquakes. Most failed dams were tailing dams or hydraulic fill
dams or small earth fill dams. Few embankments or concrete gravity dams of
significant size have been severely damaged. This good record may be largely due
to the fact that few dams have been shaken by earthquakes of duration and intensity
sufficient to jeopardize their structural integrity. It is well understood also that the
failed dams, especially tailing dams or the small earth dams were not designed or
constructed to meet rigid criteria for safety under earthquake loading.
In Table 1 a list is given of collapsed dams that have experienced significant
earthquake shaking from 1886 to 2000. It includes, where available, principal
82 Nasrat Adamo et al.

earthquake parameters, dimensions and types of dam, epicentral distances. Table 2

lists also dams which were severely damaged (but did not fail) during the same
period [4]. Examining these two tables indicates that most of the failed or the
seriously damaged dams were tailing dams and/or fill dams of generally low heights
and very few large concrete and earth fill dams. The explanation to this is already
given.
Apart from the mentioned dams (Tables 3 and 4), there were hundreds of other
dams, which had sustained minor damages or not damaged at all which may be
referred to in reference [4].
Referring back to these two tables it may be concluded that they infer the fact that
dams behave differently in response to earthquake events depending on the strength
of the earthquake on the Richter Scale (M), type of the dam and its location relative
to the epicenter of the earthquake. Moreover, Table 2 does not give full description
of damage and characteristics of the site and more details of the dams themselves.
To explain these points some important earthquakes are given hereunder with their
impacts on dams within their area of influence for clarification.

3. Important Earthquakes and their Observed Damages on

Earth fill Dams
3.1 General
Studying previous major earthquakes and the damage they left on dams serve to
illustrate the possible performance of similar dams in similar future seismic events.
The lessons learned from such case histories help designers in selecting the type of
dam best suited to the site under consideration, materials for construction and the
seismic criteria to be followed in addition to other design details.
 Dam safety and Earthquakes 83

Table 1: List of reported dams which failed as a result of earthquake

(1896- 2000)[4].
Height Earthquake Earthquake Magnitude Distance
Dam Name Country Type
[ft] name Date [km]
Augusta USA E - Charleston 13 Aug 1886 7.0 180,0
Vulcano Lake Mexico E 12 Imperial 22 June 1915 5.3 0.0
Valley
Fairmont USA E - Imperial 22 October 5.0 22.0
Valley 1916
Sheffield- 2 USA E 25 Santa Barbra 29 June 1925 6.3 11.2
Barahona Chile T 200 Talca 01 October 8.4 160,0
1928
Vulcano Lake Mexico E 12 El Centro 18 May 1940 7.1 0.0
Hosorogi Japan E 28 Fukui 28 January 1948 7.3 4.8
Coleman USA Comp - Fallon 23 August 1954 6.7 24.0
Saguspe USA E - Fallon 23 August 1954 6.7 24.0
Rogers USA M - Fallon 23 August 1954 6.7 80.0
El Soldado Chile T - Chile 28 March 1965 7.1 -
El Cobre Chile T - Chile 28 March 1965 7.1 35.0
Hayagakenuma Japan E 40 Tokachi- Oki 16 May1968 - -
Ichrigoya Japan E 26 Tokachi- Oki 16 May 1968 - -
Gamanosawa Japan E 34 Tokachi- Oki 16 May 1968 - -
Shorey Peru T - Peru 1969 - -
Huachopolca Peru T - Peru 1970 - -
Salamanca Chile T Chile 08 July 1971 7.5 110,0
Illapel Chile T 26 Chile 08 July 1971 7.5 100,0
Cerro Negro Chile T - Chile 08 July 1971 7.5 -
Mochinkoshi 1 Japan T 98 NrIzu- 14 January 1978 7.0 35
Oshima
Cerro Negro-2 Chile T 105 Chile 03 March 1985 7.7 -
Veta De Aqua Chile T - Chile 03 March 1985 7.7 -
Upper Koyoen Japan E 30 Kobe 17 January 1995 6.9 -
Central Koyoen Japan E 30 Kobe 17 January 1995 6.9 -
Niteko Japan E - Kobe 17 January 1995 6.9 <10
Shih- Kang Taiwan CG 82 Chi- Chi 17 September 7.6 0.0
1999
Legend E= Earth dam. Comp = Composite (fill/ Concrete), M= Masonry, T= Tailing, CG= Concrete Gravity
 84 Nasrat Adamo et al.

Table 2: List of reported dams which were severely damaged (but did not fail) as a
result of earthquakes [4].
Height Earthquake Distance
Dam Name Country Type Earthquake Date M
[ft.] name [km]
Ono Japan E 161 Kanto 01 September 1923 8.2 51.0
Misc. Japan E 50/8 Ojka 1939 6.6 -
Embankments
Hebgen [1] USA E 90 Hebgen Lake 17 August 1959 7.1 16.0
Hsinfengkiang Chile CGB 344 Hsinfengkiang 19 March 1962 6.1 1.1
Bella vista Chile T - Chile 28 March 1965 7.1 55.0
Koyna [1] India CG 338 Koyna 11 December 1967 6.5 3.0
Yeyuan China E 82 Bohai Gulf 18 July 1968 7.2 ?
U. Van U SA HF 80 San Fernando 09 February 1971 6.5 11.2
Norman
El Cobri Chile T - Chile 08 July 1971 7.5 80.0
Lliu Chile T - Chile 08 July 1971 7.5 -
Shimen Ling China E 147 Haicheng 04 February 1975 7.3 33.0
Touho China E 72 Tangshan 28 July 1976 7.8 -
(Douhe)
Mochinkoshi Japan T 98 Nr i- O Atssshk 15 January 1978 5.8 -
No2 [2]
La Palma Chile T 26 Chile 03 March 1985 7.7 -
Austrian [1] USA E 185 Loma Prieta 17 October 1989 7.1
Masy way [2] Luzon E 82 Philippines 16 July 1990 7.7 19.2
Niwajkumine Japan E ? Hokkaido Nans 12 July 1993 7.8 74
Lower San USA HF 125 Northridge 17 January 1994 6.7 9.4
Fernando
Lower Japan E 30 Kobe 17 January 1995 6.9 -
Koyoen
Zhong Hai China CG 82 Lijang 03 February 7.0 4.0
Legend E= Earthfill. CGB= Concrete Gravity Buttress. CG= Concrete Gravity. HF= Hydraulic Fill. T= Tailing

Major earthquakes, as natural phenomena, have in most cases left behind

destruction and life losses, which have led to a considerable number of studies and
detailed documentations. One of the important aspects so covered in these
documentations is the damages earthquakes have left on dams that were located in
their area of influence. It is of interest to note that ICOLD has devoted many of its
publications to highlight seismic action on dams together with other national
committees on large dams and government agencies.
From following case histories of dams damaged by earthquakes, but did not fail,
many serious conclusions can be derived. To follow these case histories in more
details one may refer to references [5] and [6], given in the list of references.
In the following some of the important factors contributing to damage suffered by
dams as a consequence of earthquakes, are outlined:
1. The magnitude of damage on dams is generally commensurate with the
magnitude of the earthquake. As one example; the 1923 Kanto, Japan`s
earthquake (M= 8.2) damaging the 122 feet high Ono earthfill dam. The dam
Dam safety and Earthquakes 85

was fractured in many places with one vertical fissure extending 70 feet along
the puddle clay core and settlement of about one foot in addition to
longitudinal crack on the crest of ten inches width and length up to 200 feet.
Similarly, the case of the 1985 Mexico earthquake (M=8.1) which damaged
the 197 feet earth rockfill and central clay core La Villita dam, and the 485
feet high el Infiernillo rockfill and earth core dam. Both dams experienced
considerable settlements and small permanent deformations. These damages
were added to similar damages they both had received in the previous five
events between 1975 to 1985, which had magnitudes exceeding 7.0, but
without failing. This may be compared with the 1987 Whittier Narrows (M=
6.1) earthquake, which affected many embankment dams in greater Los
Angeles area and showed no significant damages as indicated by the strong
motion records obtained from this event.
2. The mode of ground shaking during the earthquake may contribute also to the
damage produced on structures within the epicentral area, including dams. In
the 17th of October 1989 Loma Prieta event, California, USA (M= 7.1), about
dozen dams located within the epicentral area withstood the strong ground
shaking. This was because the strong phase of shaking (acceleration > 0.05g)
during the earthquake lasted less than eight seconds at rock and firm soil sites.
3. It was proved also from documenting various earthquakes that embankment
dams’ behavior during earthquakes in which sound seismic design criteria
were used, was relatively better than other dams, as they sustained lesser
damages. Good examples may be cited from the 17th September (M=7.4)
Koçaeli earthquake in Turkey. This earthquake was caused by the rupture of
the northernmost strand of the North Anatolia fault system, which produced
seven earthquakes with magnitudes more than 7.0 since 1939. None of the 48
dams located within the area were affected. Two of the dams closest to the
area where the recently completed, Yuvaçik Dam, which was about 7km of the
earthquake’s epicenter, and the Gokçe Dam, located around 55km to the
southwest close to the town Yalova.
Yuvaçik a Dam, a 108m high rock and gravel earth fill and clay core dam was first
impounded in June 1998. Its live storage capacity is 55Mm3, and its full capacity is
66 Mm3. Sound seismic design criteria were used for its design by considering
horizontal peak ground acceleration of 0.15g for the dam, and with a lower
acceleration used for design of the associated structures. The dam incurred very
little damage as a result of this event. The fact that the reservoir was not full was
fortunate as the freeboard allowed for wave action, and settlement was only 1.5m.
Total settlement before the event measured only 25mm, and following the
earthquake the maximum settlement was only 130mm, but it was reported, however,
that the tidal wave induced during the event had a height of 2. m. The earthquake
resulted in almost negligible horizontal movement of the dam of about 30mm, with
the dam recovering to half of this value within weeks from the earthquake.
The Gokçe Dam is an embankment dam 50m high and similar to Yuvaçik Dam, its
reservoir was only impounded to half of its full capacity of 25.5Mm3 at the time of
86 Nasrat Adamo et al.

the earthquake and was designed according to the same seismic criteria also. This
dam was found to have suffered no obvious damage during the event. The intake
structure is similar to that on the larger Yuvaçik Dam, and, while it showed evidence
of some cracking; it appeared generally to be very robust.
In another case, on January 7th, 1994 Northridge earthquake (M=6.7), California
USA, the earthquake induced ground motions, which were quite severe at 105 dams
located within 75km radius of its epicenter. These dams were mostly the same dams
shaken in 1971 during San Fernando (M=6.6) earthquake. Eleven earth fill and
rockfill dams experienced some cracking and slope movements as a result of the
last earthquake, yet none of them presented an immediate threat to life and property.
This satisfactory performance may be due to a large extent, from the fact that in
California, most significant dams have been reevaluated for the Maximum Credible
Earthquake (MCE), during investigations initiated after the San Fernando
Earthquake in 1971. Questionable or unsafe embankments have been upgraded or
decommissioned, or the owners have been asked to operate with partially full
reservoirs with an increased freeboard. In this connection, one of the few
embankment dams that suffered noticeable damage from the Northridge Earthquake
was the 125-foot high Lower Van Norman Dam, a hydraulic fill dam. The dam had
been abandoned as a water storage facility since 1971 San Fernando earthquake, but
was still used with empty reservoir for flood control. It experienced two to three and
a half inch-wide cracks of several hundred feet long. Some of these cracks were at
least five feet deep. Sand boils and a sinkhole were also observed along the upstream
face. Maximum crest settlement was eight inches, and maximum horizontal crest
movement was about four inches toward upstream. The 82-foot-high Upper Van
Norman Dam, which was also left with an empty reservoir since it was severely
damaged in 1971 experienced transverse cracks near its right abutment, on the
downstream slope, and near its left abutment, which were up to 60 feet long and
two to three inches wide. Maximum non-recoverable crest displacements were
about 2.4 feet of settlement and over six inches of horizontal upstream movement.
It is worth mentioning that the seismicity of the area where the two reservoirs were
located was studied by the US Geological Survey in 1974 following the San
Fernando earthquake, and the expected earthquakes` motion parameters were
obtained and included in a circular of this department [7]. In the same event the 130
foot-high Los Angeles Dam, which has replaced the two Van Norman dams, and
located between these two floods controls dry embankments, experienced
extensive, but not safety threatening cracking of its asphalt lining and settled 3.5
inches near its maximum section. Maximum horizontal crest movement was about
2.2 inches. Lastly, the Northridge Earthquake caused minor damage in the form of
transverse cracks and settlement to Lower Franklin Dam (103 feet high), Santa
Felicia Dam (213 feet high), Sycamore Canyon Dam (40 feet high), Schoolhouse
Debris Basin Dam (38 feet high), Cogs well Dam (266 feet high), Porter Estate Dam
(41 feet high), and Rubio Basin Dam (64 feet high). Adoption of sound and proper
design seismic criteria had saved all these dams from failure.
Dam safety and Earthquakes 87

4. Selection of materials of earthfill dams may also contribute greatly to their

response to earthquakes. Many earthfill dams were affected by the February
1971 San Fernando earthquake (M= 6.6) which experienced (0.15g) or greater
ground shaking. The Lower San Fernando dam, a 140 foot high hydraulic fill
dam, performed poorly and was severally damaged. The upstream face and the
crest slid into the reservoir leaving only 5 feet of freeboard against
overtopping. Lower Van Norman dam and the Upper Van Norman dam
(mentioned also above in relation to 1994 Northridge event) were seriously
damaged by this earthquake and were taken out of service as permanent
storage dams but remained to serve for flood protection dams as already
explained. The first is the hydraulic fill dam, so it experienced a widespread
liquefaction and major slope failures. Overtopping did not occur only because
the reservoir water level was relatively low when the earthquake struck. The
Upper Van Norman dam was also severely damaged.
These cases brought to the attention of engineers the potential vulnerability of
embankments constructed of poorly graded, badly compacted and saturated fine
sands and silts, and led to significant advance in the numerical methods of dynamic
analysis of dams. This also sheds light on the large number of tailing dams’ failures
during earthquakes, which result as a consequence of the uniform graded silts or
sand of which they are made of, in addition to the low degree of compaction and
the high degree of saturation of these materials making them very susceptible to
liquefaction.
In contrast to the above mentioned cases, the behavior of Kitamaya Dam in response
to the January 17, 1995 Kobe, Japan (M=6.9) earthquake merits mentioning. This
dam, an 80-foot high embankment dam was built of decomposed granite of varying
gradations with a vertical chimney drain; it was about 31km away from the epicenter
of this earthquake. At the time of the earthquake, the reservoir was at the maximum
operational level. After the earthquake, reservoir drawdown revealed a 1.0 to
1.5m scarp and bulging at the toe, so test pits were excavated into the slope to
determine the condition of the embankment at this location and to obtain samples
for testing. Beneath the riprap and gravel bedding was a loose layer of rolled
embankment, which was followed by an extremely loose layer that had contained
the slide. Unaffected embankment was found below the sliding layer, and the sliding
failure zone had a thickness of 1.5 to 2.0m. Undisturbed samples of the embankment
above the sliding block were not particularly loose. The damages suffered by the
dam were tolerable and repairable mainly because of good selection from the
construction materials and degree of compaction. In the same event, no severe
damages were observed in earth fill dams higher than 40 feet, which again was
explained by the good selection of material and high degree of control during
construction. Smaller embankment dams, however, suffered various forms of
damages such as longitudinal cracking, transverse cracking, settlement,
deformation of the dam body, and up to few complete failures; depending on their
materials and construction practices exercised. In spite of the overall assessment of
peak acceleration levels at these dams’ locations, which was estimated to be
88 Nasrat Adamo et al.

approximately 0.22g, the limited damages in this event could be partially explained
by the fact that they had good rock foundations and give support to what has been
mentioned in (3) above, in addition to a good selection of construction material, and
good control over construction.
5. In many instances chance has contributed to save many dams, or at least, has
helped to reduce damage in many dams. In these cases, the reservoirs of many
of these dams were either empty or partially full during the earthquakes, so
they were not subjected to the full hydrodynamic forces of water assumed for
the design of the dam and its ancillary works such as gates [8].
This matter, taken with the other discussed factors, helped in clarifying the
mechanism of dams’ total or partial failures or even reducing the magnitude of the
resulting damage. The example of Bhui (M=7.7), India, earthquake which struck
the Gujarat Province on January 26, 2001, can serve as a good example.
An important aspect of the Bhuj earthquake was the performance of about 300 small
and moderate size embankment dams (according to ICOLD classification) that were
constructed in this region in an effort to store water for irrigation and domestic water
supply. All these dams were hit by a/m earthquake. Fortunately, at the time of the
earthquake, being the end of the irrigation season, there was very little water in the
reservoirs and most of the damage to embankments occurred in the valley section
where the low pool kept the alluvium saturated. Following the event, the dams were
categorized based on their observed damages, downstream consequences, and their
importance to water supply. This classification was used to determine the
appropriate course of action and prioritization for restoration. Damage classes and
the total number of dams in each class are summarized in Table 3.
Multiple post-earthquake investigators noted liquefaction related damage to
embankments throughout the affected area, as evidenced by sand boils, ground
cracking and lateral spreading. Many of these dams were constructed directly on
loose alluvial deposits, and this was the reason behind the liquefaction of this
material. But the fact that those dams were mostly empty contributed to the low
hydrodynamic forces acting on them reducing damages [9].
The damages on these dams were of varying degrees of settlements, bulging of dam
body, horizontal displacements, horizontal cracks at the top and upper zones of the
dam of various depths and lengths, in addition to vertical cracks and slope slides,
but no failures were reported.
Generally Speaking, very high dams, whether zoned earth fill or rockfill dams have
responded fairly well during very strong earthquakes when they are designed
according to sound seismic criteria using suitable zoned materials. Rockfill dams,
or concrete faced rockfill dams, however, showed in many cases superior
performances over embankment dams of the same heights under the same
conditions. The following examples are of severe events and their impacts on large
modern designed, Earth fill and Rockfill dams.
Dam safety and Earthquakes 89

Table 3: Damage classifications of earthen dams resulting from Bhuj 2001

earthquakes used to prioritize repairs.

Damage Medium Minor Total

Description
Class Dams Dams Number
Significant damage. Dam is critical to
1 region and repairs will be completed 4 7 11
before monsoon
Major damage. Not possible to repair
before monsoon. Partial cut is made
2 2 0 2
to allow water to pass, but allow for
some storage
Major damage. Damage is too
3 extensive to store water. Full cut is 1 3 4
made to pass flood waters
Minor Damage. Repairs completed
4 80 165 245
before monsoon
Total 87 175 262

The Wenchuan, China, May 12, 2008 (M=8) earthquake, which had struck the 512
foot high Zipingpu concrete faced rockfill dam. This Dam was 7km from the fault
break, and it experienced an estimated peak ground acceleration of 0.5 to 0.6g. The
crest settled 3 feet damaging small parts of the face slab. Maximum settlement was
760mm. The duration of strong ground motion was up to 120 seconds at sites
underlain by deep alluvium.
In this earthquake, all the reservoirs in Sichuan and other seven nearby provinces
were damaged. There were about 300 aftershocks some of which were over M=6.
The number of earth fill dams which collapsed was 69 dam and there were 331 other
highly dangerous situations. Well-built dams, especially rockfill dams, performed
as designed [10], [11] and [12]. Nevertheless, the case of Bikou Dam, a 335 foot
high earth fill dam with central core, has to be mentioned. It had experienced an
estimated peak ground acceleration of 0.5g in this event, but its crest settled
9 inches only as a result of shaking, which is classified as mild damage [10].
Another very severe earthquake, the Mulue, Chile, February 27, 2010 (M=8.8) had
its repercussions on large dams without any of them failing. One Example is
Convento Viejo, 105 feet high, embankment dam, which was not damaged even
though it experienced a peak ground acceleration of 0.38g, which was higher than
what was expected when the dam was designed. At least 16 dams were moderately
to severely be shaken with no reported failures. Some non-threatening slope
failures, longitudinal and minor transverse cracking, however, occurred [13].
90 Nasrat Adamo et al.

In the same earthquake, the Coihueco dam a 31 meter zoned earth fill dam suffered
non-threatening sloughs on its upstream face and crest cracking apparently without
liquefaction occurring [13] and [14].
The Tohoku, Japan, March 11, 2011 earthquake (M=9.0), when more than
400 dams, which were inspected afterwards, generally performed well with minor
or moderate cracking occurring at embankment dams.
The Surikawa, a 172 foot high central core rockfill dam reacted well to this severe
event. The dam settled a maximum of 7 inches; transverse cracking of the crest
paving occurred near the abutments and the leakage temporarily increased from
18 to 25 gallons/minute. The measured peak horizontal accelerations were 0.11g at
the foundation and 0.47g at the crest. The cracks were trenched and found to be only
one foot deep.
The second high dam which was subjected to the same earthquake is the Kejauma,
which is 79 feet high, central core rockfill dam. This dam experienced peak
horizontal accelerations of 0.27 and 0.5g which was recorded at its foundation and
crest, respectively. A maximum crest settlement of 6 inches occurred and leakage
temporarily increased from 5 to 110 gallons/minute, and transverse cracking of the
crest paving were up to one inch wide. A third dam, the Minamikawa Saddle Dam,
a 64 foot high asphalt faced rockfill dam. The earthquake caused a temporary
increase in leakage from 5 to 23 gallons/minute, a crack in the asphalt face, and a
maximum crest settlement of 4 inches. A peak horizontal acceleration of 1.3g was
measured at the crest. And, 0.27g was measured on the foundation of the main dam,
1km away. The analysis of accelerations, settlements and leakage increases
measurements at these, and the other dams indicated the effects of the long duration
of the Mw 9 earthquakes shaking.
An exceptional case, however, maybe that of the Fujinuma Ike dam, a 60 foot high,
436 feet long embankment where records show an inferior performance than the
other dams. A preliminary report mentioned flaws in the embankment, such as thick
lifts, so the dam may not have been constructed to modern standards. The report
confirms the observation of long duration.

4. Important Earthquakes and their Observed Damages on

Concrete Dams
Case histories have shown that concrete dams suffered less damage from
earthquakes than embankment dams; this may be due to the nature of these
structures, the type of construction material, degree of control exercised during
construction, and to the relatively stronger foundations on which concrete dams are
normally constructed as compared to their equals of earthfill dams. Apart from Shih-
Kang dam, which failed on the 21st of September 1999, Chi- Chi Earthquake (M=
7.6) in Taiwan, no more concrete dams have failed, and this was the first case in
history so far [15], [16].
Dam safety and Earthquakes 91

It can be said that the performance of all types of concrete dams has been
satisfactory. The Shih-Kang gravity dam experience, however, confirmed that
concrete dams are vulnerable to major fault rupture.
Concrete buttresses dams when subjected to severe shaking have developed
horizontal cracks at the elevation high in dams where the downstream buttresses
intersect the vertical “chimney” section. This is an area where the stiffness of the
concrete structures significantly changes.
Major thin arch concrete dam, with a full reservoir, although have performed in a
good way, their behavior under peak ground acceleration exceeding 0.5g has yet to
be known.
Some other specific conclusions on concrete dams where damage has been
identified to indicate; there has been cracking high in the dam and where additional
features such as curbs, railings, gates, or guard and/ or control houses are located.
Cracking in buttress dams appeared to be due to upstream to downstream motions
and not cross-canyon motions. Very little in the way of increased leakage has
occurred in concrete dams subjected to major earthquakes. This can be explained,
in part, to the fact that any cracking caused by the earthquake has mainly been
horizontal and located high in the dam while the reservoir not being full in many
cases. Some rock foundations have experienced a temporary increase in seepage
following an earthquake which decreased later on.
There may be number of reasons why concrete dams have performed well and
invariably better than that predicted.
The main reasons being:
1. Concrete dams are redundant structures that provide considerable capacity to
redistribute load once damage occurs in the structure. Being so massive,
typically there is plenty of concrete volume around damaged areas of the dam
to carry loads around the damaged sections of the dam.
2. The duration of strong shaking may be too short to cause failure. Normally, it
takes considerable time at high levels of shaking to cause failure of a medium-
thick arch dam compared to a thin arch dam.
3. The dynamic tensile strength of concrete is taken as 50 percent higher than the
static tensile strength of concrete. This increase in strength makes dams
stronger during seismic shaking and increases resiliency.
4. Damping mechanisms can increase in the dam during the earthquake and
reduce the seismic impact on the body of the dam. Damping increases still as
the concrete cracks and contraction joints open and close.
5. The seismic impact of earthquake on the dam may be reduced because the
natural frequency of the dam may not match the postulated frequency content
of the earthquake taken for the design. For example, a gravity dam with a
natural frequency of 7 Hz would not align with a ground motion with a peak
spectral acceleration at 3 Hz.
6. The three-dimensional effects of the dam help prevent failure. The curvature in
plain view of the dam or the narrowness of the canyon greatly increases the
92 Nasrat Adamo et al.

seismic stability of a dam, and the potential for sliding of a gravity dam
wedged in a narrow canyon is remote.
The generally accepted potential failure modes for concrete dams during an
earthquake are cracking of concrete through the dam that forms removable blocks
and sliding of the blocks during or after the earthquake. Severely shaken concrete
dams; to date have cracked at locations of change in geometry (re-entrant corners)
but have not formed removable concrete blocks. Thus, the entire potential seismic
failure mode has not been fully achieved or experienced for concrete dams. While
concrete dams are designed to withstand a higher degree of seismic shaking than
buildings and have performed well in the past, we should not become overconfident
of their performance in the future. Great care should be taken in the design details
and quality of construction. Particular attention should be given to possible faults
located directly under the dam.
In Shih-Kang gravity dam case, the dam was hit by Chi- Chi earthquake, which was
caused from the Chelungpu faulting system rupture. Several concrete gravity and
arch dams in the same area were severely shaken during the earthquake but
performed satisfactorily. Shih- Kang dam itself could not withstand the earthquake
force along the line of the fault crossed the dam itself causing considerable
horizontal and vertical displacement, which acutely twisted the body of the dam
causing it to fail. This high gravity dam is essentially a 18-bay gated spillway. The
fault rupture extended both upstream and downstream of the dam and caused
extensive damage to bays 16 to 18 on the right side of the structure [17]. As a general
conclusion, it may be said that the performance of concrete dams has been
satisfactory [18].
Perhaps hundreds or more concrete dams of all types had been shaken by
earthquakes close to the dam sites, but only about 20 had experienced recorded or
estimated peak ground accelerations (PGA)s of 0.2g or higher. Up to 2013, some of
these dams have experienced Peak Ground Accelerations over 0.3g. The duration
of motion of the M=9.0 Tohoku Earthquake was extraordinary long lasting from
150 to 300 seconds. Following the earthquake, about 240 concrete dams were
inspected, and reports indicated that concrete dams appear to have performed very
well during the main earthquake and numerous large aftershocks. One publication
in 2017 gives a summary of concrete large dams that were shaken by a peak
horizontal ground acceleration of more than 0.3g, Table 4. Peak accelerations at the
crest were greater with full reservoirs, as expected. According to type, they were as
follows:
 Dam safety and Earthquakes 93

Table 4: Number of Concrete Dams subjected to PHGA> 0.39 g showing type and
corresponding damage [18].

Dam Type Number Damage to dam No Damage to dam Minor Damage to dam

A Gravity 10 (1 RCC) 1 5 4
B Arch 6 (1 RCC) 1 2 3
C Buttress 2 2 0 0
D Spillway 1 1 0 0
Total 19 5 7 7

To demonstrate the relatively satisfactory performance of concrete dams, few

reprehensive cases of all types are briefly described below to illustrate this point.
Examples of Concrete Gravity Dams:
1. Lower Crystal Springs Dam, this dam, a 127 foot high curved gravity dam
hit by the San Francisco 1906 (M=7.9) Earthquake, but suffered no
damage, although it was located only 0.25 miles from the San Andreas
Fault. This dam also was been moderately shaken by the 1989 Loma Prieta
(M=7.1) event without suffering damage.
2. Koyna Dam, India, this dam, a 338 foot high straight gravity dam was
shaken by the 1967 Koynanagar (M=6.7) earthquake, which occurred near
Koynanagar town close to dam site. The earthquake claimed at least 177
lives and injured over 2,200. The dam itself developed substantial
longitudinal cracking at the top. Damage was attributed to design or
construction details that would be avoided in modern structures [19].
3. Takou Dam, Japan, this is a (77m) 252 feet high straight gravity dam
completed in 2006 which underwent the Tohoku 2011 (M=9) event and the
following week (M=7.1) aftershock. In spite of the peak horizontal ground
acceleration during the main shock which was estimated at 0.40g, the dam
showed no damage except of an offset in the dam parapet wall and cracking
of the wall of the gate house. This multipurpose dam had about half full
reservoir during the main shock [20].
Examples of Concrete Arch Dams:
1. Gibraltar Dam, USA, this is a constant radius concrete arch dam 194.5 feet
(59.3m) high and 600 feet (180m) long. It was severely shaken by the 1925
Santa Barbara Earthquake (M=6.3). The estimated Peak Horizontal Ground
Acceleration (PHGA) was a greater than 0.30g. The dam; however, suffered
no damage. The dam was strengthened in 1990 by additional supporting RCC
weight on the dam downstream face, which, in effect, changed the dam from
concrete arch dam to a curved gravity dam [18].
2. Pacoima Dam, USA, this is a 372 foot high concrete arch dam. It was shaken
by the 1971 San Fernando Earthquake (M=6.6) and again in 1994 Northridge
Earthquake (M=6.7). In the first event, a peak horizontal ground acceleration
94 Nasrat Adamo et al.

of 1.25g was recorded on rock at left abutment, slightly above the dam crest.
The depth of the reservoir at the time of this event was 60% of its impounding
depth. It did not develop structural cracks or relative movements between
adjacent blocks, but the left abutment had to be strengthened using post
tensioned tendons to stabilize two large rock wedges that moved several
inches as a result of the earthquake. In the 1994 event, the dam experienced
ground acceleration also well above 1.0 g near the top of the same abutment.
During this event, the reservoir was about one-third of its impounding depth.
The dam suffered minor damages including movement of the joint between the
left abutment block by 0.5 inch, and the opening of the left end of the dam by
about two inches, while one location at the left abutment was displaced
horizontally by 19 inches and a rock mass was displaced by 14 vertically in
another location.
3. Rapel Dam, Chile, this is a double curvature arch dam 364 feet high and 886
feet long that was hit by two earthquakes. The first was the 1968 Santiago
earthquake (M=7.8), and the second was the 2010 Maule (M=8.8) earthquake.
Prior to the first event, a swarm of 300 earthquakes of lesser magnitude
occurred. Measured peak free-field accelerations near the dam were 0.31g in
the cross- canyon direction, 0.14g in the upstream to downstream direction,
and 0.11g vertical. The arch dam did not experience any damage, but the
appurtenant structures did have some damages. The spillway walls were
cracked and there was leakage through the wall of the right spillway. The
upper part of one intake tower cracked and separated from the dam. In the
2010 Maule earthquake, the reservoir was full, and measurements showed that
the peak horizontal acceleration at the site was 0.3g. One concrete block of the
dam at the left abutment of the dam which was next to the fault “Nido de
Aquila” showed a rise of 0.02 inches. Seepage again increased along the right
abutment; this time from normal 3.4gal/sec to 10.6gal/sec, and some concrete
pavement at the dam crest cracked.
Examples of Concrete Buttress Dams:
1. Hsinfengkiang Dam, China, this is a 344 foot high buttress dam, which was
shaken by the nearby earthquake of (M=6.1) in 1962. The dam developed
substantial longitudinal cracking near the top, but damage was attributed to
design and construction details that were avoided in modern structures.
2. Sefid Rud Dam, Iran, this is a 388 foot high buttress dam which was hit by the
1999 Manjil Earthquakes (M=7.4). It suffered severe cracking in the upper
part of some buttresses and other forms of damages. It was rehabilitated and
remained in services. Some 20 years later, a blister on the steel lining in the
elbow of one of the two morning glory spillways in the left abutment
appeared, and the increased leakage led into the discovery of previously
unnoticed damage caused by the previous earthquake [17].
Dam safety and Earthquakes 95

Examples of Rolled Compacted Concrete (RCC) Dams:

1. Shapai Dam, China. This is a 132m high RCC arch dam which was
moderately shaken by the Wenchuan (M=8) in 2008. The dam withstood this
event without damage to its body.
2. Miyatoko Dam, Japan. This is a 157 foot high RCC dam which was hit by the
Tohoku (M= 9.0) earthquake of 11 March, 2011. The dam was located
approximately 13km north of Sendai in Miyagi, area where the peak horizontal
ground acceleration was greater than 0.7g. A strong motion instrument located
in the gallery recorded peak horizontal ground acceleration of 0.32g during
this earthquake. No damage was recorded for Miyatoko and Takou Dams, and
possibly others in Japan, will need to be confirmed at a later date [17].

5. Seismic Design Criteria for Large Dams

Large concrete dams were among the first structures for which seismic analysis and
design had been performed. The seismic analysis method that was originally
developed by Westergaard in the 1930s for the Hoover Dam has found worldwide
acceptance among designers of concrete dams (Westergaard, 1933) [21]. This
relatively simple pseudo-static analysis method accounts for both the inertial effects
of the dam body and the hydrodynamic pressure acting on the vertical upstream face
of a dam. It was a common practice to use a seismic coefficient of 0.1,
corresponding to a horizontal force equals to the weight of the dam times a ground
acceleration of 0.1g. The USBR, depending upon the size of the dam and the seismic
risk, however, used seismic force which was assumed to range from 0.05 to 0.15
times the weight of the structure. For larger dams, the bureau combined horizontal
acceleration effects with a vertical component, which was 50 percent of the
horizontal acceleration; the assumed directions of the two components were those
most unfavorable to structural stability. Most large dams design in countries other
than USA adopted similar criteria. For example, Bhakra Dam in India, a 740-foot-
high concrete gravity structure located about 180km from the epicenter of the
Richter magnitude (M=8.6) Kangra earthquake of 1905, was designed for a lateral
force coefficient of 0.15 and a vertical force coefficient of 0.075. The U.S. Army
Corps of Engineers practice required the use of seismic coefficients for sliding and
stability analyses of concrete dams and structures. Hydrodynamic pressures also
were considered by similar methods in some cases [22]. The pseudo-static analysis
method, however, proved in many cases to be unsatisfactory as it did not meet many
seismic modes of ground shaking and ground acceleration. This method has
therefore been replaced by the Dynamic Response Analysis, especially for large
dams. Today, the seismic safety of dams is assessed based on typical failure modes
and the inelastic deformations of dams due to ground shaking, and the evaluation of
earthquakes. In this approach, the seismic design criteria for large dams defines
many levels of safety, as per the magnitude of the earthquake obtained from the
seismic history of the area, and depending on the importance of the dam or related
structures, as follows:
96 Nasrat Adamo et al.

Maximum Credible Earthquake (MCE): It is the event which produces the largest
ground motion expected at the dam site on the basis of the seismic history and the
seismo-tectonic setup in the region. It is estimated based on deterministic
earthquake scenarios. According to ICOLD, the ground motion parameters of the
MCE shall be taken as the 84 percentiles (mean + one standard deviation).

Maximum Design Earthquake (MDE): For large dams, the return period of the
(MDE) is taken as 10,000 years. For dams with small and/ or limited damage
potential, shorter return periods can be specified. The (MDE) ground motion
parameters are estimated based on a probabilistic seismic hazard analysis (PSHA).
According to ICOLD, the mean values of the ground motion parameters of the
(MDE) shall be taken. In the case where a single seismic source (fault) contributes
mainly to the seismic hazard, uniform hazard spectra can be used for the seismic
design. Otherwise, based on the disaggregation of the seismic hazard (magnitude
versus focal distance) different scenario earthquakes may be defined.

Safety Evaluation Earthquake (SEE): The SEE is the earthquake ground motion
which a dam must be able to resist without uncontrolled release of the reservoir. For
major dams, the SEE can be taken either as the (MCE) or (MDE) ground motions.
Usually, the most unfavorable ground motion parameters have to be taken. If it is
not possible to make a realistic assessment of the (MCE); then the (SEE) shall be at
least equal to the (MDE). The (SEE) is the governing earthquake ground motion for
the safety assessment and seismic design of the dam and safety-relevant
components, which have to be functioning after the (SEE).

Design Basis Earthquake (DBE): The (DBE) with a return period of 475 years is
the reference design earthquake for the appurtenant structures. The (DBE) ground
motion parameters are estimated based on a (PSHA). The mean values of the ground
motion parameters of the (DBE) can be taken. Noting that the return period of the
(DBE) may be determined in accordance with the earthquake codes and regulations
for buildings and bridges in the project region.

Operating Basis Earthquake (OBE): The (OBE) may be expected to occur during
the lifetime of the dam. No damage and/ or loss of service must happen. It has a
probability of occurrence of about 50 % during the service life of 100 years. The
return period is taken as 145 years. The (OBE) ground motion parameters are
estimated based on a (PSHA). The mean values of the ground motion parameters of
the (OBE) can be taken.

Construction Earthquake (CE): The (CE) is to be used for the design of temporary
structures; such as coffer dams and considers the service life of the temporary
structure. There are different methods to calculate this design earthquake. For the
temporary diversion facilities, a probability of exceedance of 10% is assumed for
the design life span of the diversion facilities. Alternatively, the return period of the
Dam safety and Earthquakes 97

(CE) of the diversion facilities may be taken as that of the design flood of the river
diversion. The MDE, DBE, OBE and CE ground motion parameters are usually
determined by a probabilistic approach (mean values of ground motion parameters
are recommended), while for the (MCE) ground motion deterministic earthquake
scenarios are used (84 percentile values of ground motion parameters shall be used).
However, for the MDE, DBE, OBE and CE also deterministic scenarios may be
defined.
If reservoir-triggered seismicity (RTS) is possible; then the (DBE) and (OBE)
ground motion parameters should cover those from the critical and most likely
(RTS) scenarios as such events are like to occur within years after the start of the
impounding of the reservoir.

The different design earthquakes are characterized by the following seismic

Parameters:
1. Peak ground acceleration (PGA) of horizontal and vertical earthquake
components.
2. Acceleration response spectra of horizontal and vertical earthquake
components typically for 5% damping, i.e. uniform hazard spectra for (CE),
(OBE), (DBE) and (MDE) obtained from the probabilistic seismic hazard
analysis (mean values), and 84 percentile values of acceleration spectra for
(MCE) obtained from the deterministic analysis using different attenuation
models.
3. Spectrum-compatible acceleration time histories for the horizontal and vertical
components of the (MCE) ground motion determined either from a random
process or by scaling of recorded earthquake ground motions. The artificially
generated acceleration time histories of the horizontal and vertical earthquake
components shall be stochastically independent. To account for aftershocks, it
is recommended to increase the duration of strong ground shaking.
In case of fault movements, similar estimates are required for the ground shaking.
It appears that it is quite difficult for dam designers to get quantitative estimates of
fault movements for the different types of design earthquakes. So, they are justified
in such cases to use simplified load and analysis models that lead to a safe design,
even if the load model does not comply fully with the real nature of the earthquake
ground motion.
Considering the Appurtenant Structures of dams, these may be assigned to the safety
level to which they belong with respect to the safety of the whole dam as follows:

Safety class 1: All elements related to the safe control of the reservoir, i.e. Bottom
outlets and spillways are defined as safety-critical or safety relevant elements; they
shall be designed for (SEE) and OBE (serviceability).

Safety class 2: All structures and components related to power production

(penstock, power intake, powerhouse, tunnels, caverns, turbines, switchyard,
transmission lines etc.), water supply, irrigation, navigation etc.; they shall be
98 Nasrat Adamo et al.

designed for (DBE) with high importance factor or according to earthquake building
code as minimum requirement for the region under consideration.

Safety class 3: Other items which can easily be replaced/repaired when damaged
and whose failure has acceptable consequences; they shall be designed for (DBE)
or according to earthquake building code, [23], [24], [25] and [26].

6. Seismic Hazards and Their Impacts on Dam Safety

During the planning and design stage, safety of the dam may be ensured as far as
seismicity is concerned by concentrating the study on certain seismic factors, which
are the following:

Proximity of Site to fault Lines

The selection of the most appropriate type of dam may be affected by its distance
from a known fault line. In a known fault zone, the possibility of existence of some
branches from the main fault line should be given serious consideration. The Shih-
Kang concrete gravity dam (Barrage) already discussed happened to be located on
a branch of a main fault which had spectacular horizontal and vertical displacements
during the 1999 Taiwan earthquake that resulted from the rupture of this fault
causing failure of the dam. It is questionable; however, if any other type of dam
could have endured this severe displacement, which had occurred in this case.
Avoiding altogether such site for building a dam would have been the safest solution
if the existence of such a fault was confirmed beforehand. Fault movement in the
dam foundation or discontinuities in dam foundation near major faults can be
activated, causing structural distortions [27].
In the case of Rubar Lorestan Dam at the Zagros Mountains in western Iran, the site
was only 1.6km from the Saravand- Baznavid Fault. The main recent fault is thought
to have moved about 50km in the last 3-5 million years implying a horizontal slip
rate of 10–17mm/yr, and it was the source of frequent earthquakes of M=6 to M=7
earthquakes. The peak ground acceleration at the dam site was calculated using
seven different attenuation formulae, and it varied from 0.52g to 0.61g. After long
discussions as to whether a rockfill dam might be better able to resist fault
movement than the rolled compacted (RCC) dam, but possibly it would be much
more expensive than RCC dam as it would need spillway tunnels through the
abutments whereas spilling water could have been routed over an RCC dam at little
extra cost. The decision was to build the cheaper RCC dam even that the Rockfill
dam is a little safer [28].
Dam safety and Earthquakes 99

Liquefaction
Liquefaction is a phenomenon in which the strength and stiffness of a soil are
reduced by earthquake shaking or other rapid loading. Liquefaction and related
phenomena have been responsible for tremendous amounts of damages in historical
earthquakes around the world.
Liquefaction occurs in saturated soils, that are soils in which the space between
individual particles is completely filled with water. This water exerts a pressure on
the soil particles that influences how tightly the particles themselves are pressed
together. Prior to an earthquake, the water pressure may be relatively low. However,
earthquake shaking can cause the water pressure to increase to the point where the
soil particles can readily move with respect to each other. The previous concepts
apply to any earth retaining structures that is in contact with water and located in
seismic regions, such as an earth fill dam. Liquefaction, therefore, is a serious
potential problem for dams built on or with low density, saturated sands. In such
cases, liquefaction potential may exist either in the embankment material itself, or
in the foundation alluvial (in case the dam is resting on soil), and/or the abutment.
The crest of the 40m high Lower San Fernando Dam settled 8.5 meters in the 1971
earthquake which had a magnitude of 6.6. The dam was built of hydraulic fill, which
is particularly vulnerable to liquefaction, because of the low density of the fill.
Fortunately, the water level was about 11m below the crest before the earthquake,
but only 1.5m of badly cracked material remained after the event. 80,000 people
living downstream of the dam had to be evacuated. The 8m high Sheffield dam
failed completely in the magnitude 6.3 Santa Barbara event of 29 June, 1925. The
dam and its foundation were silty sand, and some experts have blamed the failure
on liquefaction of these materials. Krasnodar Dam in Russia near to the Black Sea
is 11.5km long and built of hydraulic fill. It holds 2,914Mm3 in a reservoir with an
area of 413km2. A seismic study was carried out by Swiss Experts who
recommended improved drainage at a cost of 56 million USD. The cost of failure,
if happened, would have cost an estimated 3 billion USD at the time (about 2000).
Tailings Dams have a particularly bad record with hundreds killed in various
liquefaction failures in Chile. One example from the Dominican Republic is the
84m high Las Mejitas tailing dam, which holds 48 million tons of very acidic
tailings at the Pueblo Viejo gold mine, only 35km from the Septentrional-Orient
Fault Zone (SOFZ), where the estimated (MCE) at the site is 0.5g. It was reported
that there was lack of adequate zoning in the dam and further studies were
recommended for the dam in 2002.
Under seismic action, liquefaction of the tailings which would be denser than water,
seems to be possible, [28], [29], [30] and [31].

Seiches
Some authors consider that reservoir oscillation due to ground shaking, and the
following seiches are of lesser importance on dam safety than cases of ground
shaking itself, which may cause vibrations in the dams and its appurtenant structures
and equipment, or from fault movements in the dam foundation causing structural
100 Nasrat Adamo et al.

distortions, and also fault displacement in the reservoir bottom causing water waves
or loss of freeboard; or even mass movements into the reservoir resulting from
landslides causing impulse waves in the reservoir. Some other authors, on the hand,
give examples of some serious cases of seiche which might have been caused by
resonance of water in reservoirs that were disturbed by seismic activity acting on
dams, [28].
A preliminary risk analysis was completed recently of two dams in Japan. One of
them, a 131m high dam, settled 30mm in an M=7 earthquake in 1961. The
probability of failure of the two dams as a direct result of an earthquake was shown
to be negligible even though a seismo-genic fault passed only 250m from the right
abutment of one of them. Moreover, the analysis indicated that the chance of seismic
seiches and seismically induced landslides into the reservoir causing displacement
wave and overtopping of the dam was also low. The calculated probability of dam
failure from seismic seiches has been checked with previous dam failure statistics,
and it looked that such probability was very low as compared to other modes of
failures; as seen from Table 5.

Table 5: Estimated annual probabilities for the principal modes of failure.

Failure Mode Annual Probability Return Period (years)

Internal erosion 11.3 x 10-6 88,496

Overtopping in typhoon 50 x 10-6 20,000
Seismic seiches 4.3 x 10-6 232,558
Landslide 9.3 x 10-6 107,527
M & E spillway gate 13.2 x 10-6 75,758
TOTAL 88.1 x 10-6 11,350

Two cases of near failure as result of seiches are the Yuvaçik Dam, Turkey, in the
1999 earthquake and the Hebgen concrete-cored earth fill dam in Montana in the
1959 earthquakes. In the first case, the amplitude of the seiche was about 5m but
the reservoir was not full at the time so there was no overtopping. If the reservoir
had been full there could have been up to 1.68m overtopping at the abutments and
0.25m at the center of the dam. In the 35m high Hebgen dam, the reservoir was full
at the time. This seiche was caused by fault movements crossing the reservoir rather
than by ground shaking. A few minutes after the first shock, the caretaker had
rushed to the dam and, in the moonlight observed the reservoir action from the high
ground above the right abutment. The first waves had already overtopped the dam
before he arrived. A few minutes later, another wave struck the dam with such
momentum that water one-meter-deep ran uniformly over the crest for 10 minutes.
Subsequently, the wave receded and seemed to travel to the other end of the
reservoir. After 10 minutes it returned, and water flowed over the crest for another
10 minutes’ period. This action was repeated and although the estimates of depths
Dam safety and Earthquakes 101

and durations are approximate, there can be no doubt that the water flowed over the
dam at least four times causing some damage, but the dam did not fail. The
magnitude of the earthquake was 7.5 to 7.8 with one of the main faults passing
within 215m of the dam.

Landslide
Landslides can form real hazard to populations in mountainous areas, where
relatively unstable masses of rock and earth may slide down in mass and hit
communities living below causing real havoc by destroying houses and structures
and burying people. Triggering of these landslides may be initiated by rainfall
causing saturation of the soil mass and its sliding over the surface of a weak surface
within. Such failure can also happen as a consequence of seismic ground shacking
resulting from an earthquake, especially when such weak potential plain of failure
exists [33] and [34]. If such landslide occurs into the reservoir of a dam, then there
is a chance that overtopping, and failure of the dam will follow causing destruction
and human fatalities downstream. Details and summary of landslides associated
with dams and resulting fatalities are given by Professor David Petley from Durham
University in a presentation given in the 2013 through an International Conference
held in Padua in 2013 on Vajont (1963) incident entitled “Landslides and Large
Dams”. The presentation examines losses associated with large dams in the last few
years and shows in particular, that there have been over 500 deaths in landslide-
related accidents during dam projects [35]. This presentation was followed after
about one month by a blog written by the same author giving some more details
[36]. In a paper entitled “Global Losses from Land Slides Associated with Dams
and Reservoirs”, it is stated also that the risk of landslides can be exasperated by the
seismic activity resulting from impounding large reservoirs known as Reservoir
Induced Earthquakes (RTS). From the technical perspective, it may be said that
seismic activity can profoundly alter the rates of activity of the landside, such that
conditions that apply during a site investigation phase may no longer be current later
on [37]. Research work, however, has also shown that seismically triggered
landslides can affect dams. One particular study indicated that the behavior of an
earth dam may depend on the probability of landslide slipping in its reservoir.
Despite previous studies which have been done on the landslides without dynamic
analysis, this study has investigated the effect of landslide impacting on earth dam
during an earthquake. It showed that safety factor during an earthquake for earth
dams decreased approximately by 40 percent, and the probability of landslide and
its undesirable effects on the earth dams in the presence of reservoir water increased
significantly [38]. Another study has concluded that moderate to large earthquakes
can trigger landslides, and these landslides commonly cause a significant proportion
of total earthquake damage. The ability to predict slope stability during earthquakes
is especially important for seismic hazard analysis of dam projects [39].
Earthquake hazard is multi- sided for large storage dams. Recent earthquakes have
demonstrated this in many such projects, which have been affected in many
102 Nasrat Adamo et al.

different ways, but those which are related to landslides alone can be listed among
these, as in the following:
1. Rockfalls causing damage to gates, spillway piers cracking, retaining walls
overturning, surface powerhouses cracking and puncturing, and various
damages to electro-mechanical equipment, penstocks, switchyards,
transmission lines, etc.
2. Mass movements of landslides and rockfalls into the reservoir causing impulse
waves and overtopping of dams.
3. Mass movements blocking rivers and forming landslide dams and lakes whose
failure may lead to overtopping of downstream run-of-river power plants or the
inundation of powerhouses and the electro-mechanical equipment.
4. Mass movements blocking access roads to dam sites and appurtenant structures
[40].
Landslide problem in relation to dam safety has been extensively studied on selected
case histories treating it according to, landslide type, rock fall; including talus
deposits, problems encountered due to interaction between dams and landslides, and
mitigation measures to stabilize such as landslide.
USGS in its professional paper “Interaction of Dams and Landslides-Case Studies
and Mitigation”, published in 2006 presented very good information on landslides
[41]. Although this publication does not treat seismic triggered landslides with
respect to dam safety, but it gives extensive general information on landslides in
relation to dams as such.
Generally, it can be said from the foregoing that the probability of dams’ failure
resulting from seismically triggered landslides is very remote. Failure probability
of dams caused by landslides is shown to be 9.3 x 10-6 as already indicated from
Table 5. If this probability is coupled with the probability of occurrence of an
earthquake which is strong enough to cause such a landslide, then the resulting
probability of failure of the dam becomes very remote indeed.

7. Reservoir Triggered Seismicity (RTS)

One important issue associated with the question of dams’ safety is the phenomenon
of Reservoir Triggered Seismicity (RTS). This phenomenon is not fully-understood,
but basically what happens is this; when a dam is built and the reservoir filled
with water, the amount of pressure exerted on the earth in that area changes
dramatically. When the water level of a reservoir is raised, the pressure on the
underlying ground increases, and when the water level is lowered, the pressure
decreases. This fluctuation can stress the delicate balance between tectonic plates
in faults beneath the surface, possibly causing these plates to shift. Another factor
is the water itself; when the water pressure increases more of it is forced into the
ground, accordingly, filling cracks and crevices. This water pressure can expand
those cracks and even create new tiny ones in the rock causing greater instability
below ground. What is more is, as the water sinks deeper, it can act as sort of a
Dam safety and Earthquakes 103

lubricant for rock plates that are being held in place by friction alone. The
lubrication can cause those plates to slip [42].
The sure fact is the dam cannot cause an earthquake all by itself. The risk factors,
specifically unstable fault lines, have to be there already. With the right conditions
in place, a dam can trigger the event earlier than would have happened naturally,
and perhaps even increases its magnitude, which is why it is so dangerous to build
a dam over a known fault.
A Study case of (RTS) may be cited here, as an illustration; this is the case of
Danjiangkou reservoir. The study explains the various factors contributing to trigger
the seismicity which had begun after impounding this reservoir in 1967 in Henan
and Hubei provinces in central China. The Danjiangkou dam was built in 1958, and
impoundment of the 174.5 × 108 m3 capacity reservoir began on 5 November 1967.
The originally designed height of the dam was 97m, making the impounded
reservoir one of the larger man-made lakes in Asia. As the source of Middle Route
of South-to-North water diversion in China, the height of the dam was supposed to
be increased to 111.6m starting in 2005 to be completed in 2010. Tectonically, there
are four major Paleozoic deep regional crustal faults that crisscross in the reservoir
area. They are overlain by about 4 to 5km thick sedimentary cover. Below the
sedimentary cover, the crustal thickness varies from 33 to 34 km in the east, to about
37 to 41km in the mountainous area to the west. The deep crustal faults in the
reservoir area are: Danjiang, Junyun, and Gonglu and Hanjiang faults. The study
confirms that this tectonic setting is responsible for the (RTS) induced in the area.
The increased load of the reservoir has caused compressive stresses on the bottom
of the reservoir which can destabilize the underlying faults, and form approximately
150KPa tensile stresses in horizontal direction on the surface within the periphery
of the reservoir, which might help to open small shallow fractures and promote the
permeation of water into deeper rocks. This suggests that the induced seismicity in
the reservoir area is mainly attributable to water and migration along the Danjiang
and Junyun faults [43]. Other studies have also shown that the reservoir triggered
an earthquake is linked to dams higher than about 100m, to large reservoirs
(capacity greater than 500 x 106m3), rate of reservoir filling, and to new dams of
smaller size located in tectonically sensitive areas. This means that the causative
fault is already near to failure conditions and so the added weight stresses and pore
pressures propagation due to reservoir impounding, can trigger the seismic energy
release.
1. The detection of the reservoir induced seismicity may be performed in two
phases: The first phase includes study of historical seismicity and surveys of
the reservoir and surrounding geological structures, aiming at identification of
possible active faults.
2. Second phase is carried out, starting at least one or two years prior to
impounding, with the installation of a permanent network of seismometers and
precise levelling beacons and use of instrumentation to detect active fault
movements, in addition to carrying out reservoir slope stability studies.
 104 Nasrat Adamo et al.

In Table 6 some examples of dam sites are presented, where induced earthquakes
with magnitude higher than 5 on the Richter scale have occurred.
The reservoir triggered an earthquake differs from the natural earthquake as those
are likely to occur, if ever, within the first 5 years after the filling of the reservoir or
when the reservoir reaches the maximum level. These earthquakes have a shallow
focus, and their epicenters are close to the dam sites or reservoirs [44].

Table 6: Examples of dams with induced seismicity [44].

Reservoir Induced
Year of seismicity Prior
Height volume
Dam Country Type impounding seismicity
(m) (x 106 m3) M year
Marathon Greece Gravity 63 41 1930 5 1938 moderate
Hoover U.S.A. Arch gravity 221 36703 1936 5 1939 ---
Kariba Zimbabwe/
Arch 128 160368 1959 5.8 1963 low
Zambia
Haifengkiang China Buttress 105 10500 1959 6.1 1962 aseismic
Koyna India Gravity 103 2708 1964 6,5 1967 low
Kremasta Greece Embankment 165 4750 1965 6.3 1966 moderate
Roi Greece Embankment 96 1000 1969 6.3 moderate
Constantine
Oroville U.S.A. Embankment 236 4298 1967 5.7 1975 moderate
Nurek Tajikistan Embankment 330 11000 1972 5 1977 moderate
Tarbella Pakistan Embankment 143 14300 1974 5,8 1996 low
Aswan Egypt Embankment 111 163000 1974 5.3 1981 aseismic
Polyphyton Greece Embankment 112 2244 1974 6.7 1995 aseismic
Morns Greece Embankment 126 640 1961 ---- ---- aseismic

8. Seismic Instrumentation for Dams

Large dams are particularly sensitive to earthquakes. They are usually built in
valleys, which exist because active erosion is taking place. This implies there has
been recent uplift produced under compressional tectonic force reverses or thrust
faults, so reverse or thrust faults dip under the up- thrown block. Many dams,
therefore, have an active fault dipping under them. Reservoirs created by dams can
also trigger an earthquake if certain conditions are met, i.e.; if the reservoir is deep
enough and the water load is appreciable, the ground underneath is saturated and
hence the high groundwater pore pressure decreases the effective strength of the
rock under the reservoir, the area is already under considerable tectonic stress, and
faulting exists, and in many cases, this activity is increased by raising and lowering
of the water level. Such reservoir triggered seismicity will have its repercussions on
the dam and its appurtenant structures in the same way as natural earthquakes [45].
It is of vital importance that, seismic events whichever their source, must be studied
and their impacts on dams analyzed. Any meaningful dynamic analysis of dams
which reliably predicts the behavior of dams during very strong ground shaking
requires sufficient data on the spatial variation of ground motion acting on the dam’s
Dam safety and Earthquakes 105

body, and such data may be collected by seismic instrumentation. In such a capacity
effective seismic instrumentation and monitoring of dams can help to monitor the
safety of dam and its proper maintenance and up-keep. It can also considerably
contribute to the overall activities for seismic risk reduction and facilitates response
studies that lead to improved understanding of the dynamic behavior and potential
for damage to structures under seismic loading. It is necessary; therefore, that
responsible dam personnel and higher staff have proper understanding of the work
of instrumentation, their maintenance and relevant records/data and their analysis.
The specific purposes of dams seismic monitoring may be summarized by the
following:
1. Determination of the location of earthquake epicenters relative to dam location
and their depth.
2. Definition of main earthquake parameters: magnitude, frequency
characteristics and some indications of focal mechanisms.
3. Prediction of the mode of occurrence of future earthquakes.
4. Provision of data on the dynamic behavior of the dam body for the purpose of
objective evaluation of its functioning immediately after the occurred
earthquake.
5. Verification of design parameters by the actual behavior of the dam body
under an earthquake.
To achieve the aforementioned goals, it is necessary to investigate and monitor the
dam site by means of various seismic instruments, so it is desirable to install seismic
instruments on the dam itself and at near locations for this purpose. The various
types of instruments required are:
1. Strong-motion accelerographs for recording potentially destructive ground
shaking and resulting dam vibrations.
2. Sensitive seismographs for determining the local seismicity. A minimum of
two strong-motion accelerographs should be installed on the dam and a
minimum of two should be installed in the immediate vicinity of the dam.
Each accelerograph should record the three components of motion, should
have a natural frequency of approx. 20Hz, a recording speed of approximately
1cm/s. The sensitive seismographs are intended to record the local seismicity
in the vicinity of the dam site before construction, and to detect any changes in
seismicity during reservoir filling. A vertical component seismometer (1Hz-
5Hz) with visual recorder and approximately 10,000 magnifications at 1Hz is
recommended.
3. Seismic data analysis system.
4. Magnetic tape play-back system.
5. Digital time marking system.
6. Digital event recorder.
106 Nasrat Adamo et al.

The instruments shall be suitable to record the earthquake detail up to magnitude

8.0 on Richter`s scale. The microearthquake may contain frequencies up to
hundreds of Hertz (Hz), while an earthquake recorded in a seismic net may contain
frequencies between 1 and 20Hz. So, it is very difficult to achieve such a dynamic
range in a single recording system with an equal response to all the frequencies,
[46]. The sketch in Figure 3 indicates recommended dam seismic instrumentations
and their arrangement; at the crest, abutments, foundations of the dam and free field
in the vicinity. In this way, full information on the ground shaking at the site in
addition to a full dynamic response of the dam may be obtained and then be
correlated with other available data from seismic national or international networks
[45]. Safer dams with respect to earthquakes can be designed and built now with
more confidence. Technologies for designing and constructing dams and
appurtenant structures that can safely resist the effects of strong ground shaking are
available. This includes in addition to developed seismic criteria and analytic
methods of design, an array of modern seismic instrumentation as a good asset to
analyze the dynamic response of dams during earthquakes, to learn more on their
behavior in such events and draw more lessons for future works.

Figure 3: Recommended dam seismic instrumentation [45].

Dam safety and Earthquakes 107

9. Inspection Procedures after Seismic Events

In order to assess the safety of dams affected by earthquakes, thorough investigation
is usually required by the dam safety authorities immediately after the event.
According to the American Army Corps of Engineers (USACE), this assessment is
to be done in certain recommended procedure, which is done by carrying out initial
inspection and evaluation survey, followed by Post-Earthquake Inspection
(Engineering).
Inspections and Evaluation Surveys
The design of dams for earthquake loading is hampered by the infrequent
opportunity to compare actual performance with the design assumptions.
Significant damage is unlikely to occur if peak ground acceleration (PGA) is below
0.1g; nevertheless, structural integrity cannot be assumed if earthquake loadings
occur below the design criterion. For example, earthquakes have occurred in places
where significant seismic activity was not expected, and where an earthquake
induced loads may therefore have not been adequately considered in the design of
older structures.
Many types of structural damages can be induced by ground motion from
earthquakes. Any post-earthquake change in appearance or functional capability of
the dam should be evaluated and reported. The inspection engineers should look
carefully into any such damage, whatever small it may be, and report any of the
following:
1. Symptoms of induced stresses evident by cracks in concrete dams or earth
embankments.
2. Misalignment of hydraulic control structures or gates, and possible loss of
function.
3. Cracks in service buildings or pavements.
4. Rock falls from the adjacent slopes on the dam.
5. Loss of freeboard induced by dynamic loading on earth dams or levees, which
may result from settlement.
6. Dynamic loading on earth dams or levees may also cause localized
liquefaction within the embankment sections or earth foundations leading to
deformation of the embankment.
7. Hydrologic changes also can indicate the potential for subsurface deformation,
such as abrupt changes in static water levels in wells and increased turbidity.
8. The appearance of surface seeps can indicate the formation of new seepage
paths within the foundation or embankment section.
9. Ground motion-induced landslides may occur in susceptible areas of reservoir
rims, causing embankment overtopping by waves and serious damage.
Inspection and Evaluation Programs
The following characteristics can be useful for prioritizing the sequence of structural
inspections and damage evaluation:
108 Nasrat Adamo et al.

1. Significant Damage to Structures in the Vicinity: If the project is located

within the vicinity of an earthquake that results in significant structural
damage, an engineering evaluation and inspection should be carried out by a
component team from the local authority, while the project operations`
personnel will conduct and document an initial damage survey inspection of
the project without delay. The project’s emergency operations plan should be
invoked. Actions during emergency operations should include consideration
of the effects of potential aftershocks.
2. No Significant Damage to Structures in the Vicinity: If an earthquake is felt
at the project during normal working hours but causes no or insignificant
damage to structures in the vicinity, project operation personnel should make
an immediate inspection. If the event occurs after normal working hours, the
project operations manager will determine the need for the immediacy of the
inspection but in no case will it be delayed beyond the beginning of the
following day. This inspection should determine and document, (i) whether
there is evidence of earthquake damage or disturbance, and (ii) whether
seismic instrumentation, if present, has detected ground motion, then a Safety
Emergency Alert should be issued and follow reporting procedures.
Combinations of Magnitude and Distance
1. The following criteria presented by USACE and known by “Earthquake
Inspection Criteria Simplified Model” are listed below and illustrated by
Figure 4 allow operations managers to assess the immediate need for site
inspections and evaluations based on earthquake magnitude and proximity to
an earthquake epicenter. If the magnitude and distance fall within the specified
ranges, project operation personnel will conduct an inspection and check all
the seismic instruments and alarms, if any are present at the project. If no
seismic instruments are located at the project, then earthquake magnitude and
epicenter data can be obtained from regional or international seismological
centers such as; The Global Seismographic Network (GSN), The International
Seismological Centre (ISC) or The European Mediterranean Seismological
Centre (EMSC).
2. If any of the seismic alarms indicate a (PGA) of 0.05g or greater, or if damage
is identified, then the inspection results performed by the project staff will be
reported immediately to the designated personnel within the next higher
engineering authority who will carry out themselves a more comprehensive
inspection. The USACE criteria, aforementioned, for the need or otherwise to
carry out the inspection is stipulated in the following ranges and illustrated in
Figure 4:
- Magnitude 4.5 through 4.9 within 16 km (10 mi).
- Magnitude 5.0 through 5.9 within 80 km (50 mi).
- Magnitude 6.0 through 6.9 within 120 km (75 mi).
Dam safety and Earthquakes 109

- Magnitude 7.0 through 7.9 within 200 km (125 mi).

- Magnitude 8.0 or greater within 320 km (200 mi).

Figure 4: Earthquake Inspection Criteria – Simplified Model [47].

Analysis of Results and Final Report

It is of utmost importance that all compiled inspection reports be studied carefully
and analyzed in the light of, but not limited to:
1. The physical characteristics of the earthquake event, including the magnitude
of the earthquake, the distance between the epicenter and the project, the peak
horizontal ground acceleration (either calculated or measured at the site).
2. The condition of the structure or embankment as documented during
inspections.
3. The structure or embankment performance history. Whenever feasible, on-site
instrumentation data is incorporated into the post-earthquake safety evaluation
programs to fulfill inspection.
This Analysis may indicate the need to install in the future additional special types
110 Nasrat Adamo et al.

of instrumentation to be incorporated in selected structures to improve measurement

of forces, pressures, loads, stresses, strains, displacements, deflections, or other
conditions relating to damaging and structural safety and stability in case of an
earthquake.
The final report shall not fail to include also the results of inspections of features
not visible on the surface, such as conduits, tunnels, and galleries, which should
have been performed with particular attention as damage or differential movement
of these features could lead to delayed adverse effects. Moreover, inspections of
underwater project features should be performed to include positional surveys of
structures and/or channels as deemed necessary.
Summaries of the various instrumentation data and those of accelerometers or other
types of strong motion instruments, if they have been installed, together with other
observations’ data for each inspection should be also appended to the final report
for permanent record and reference purposes.
This final report must be viewed as of great importance and value as it will form the
basis for major remedial work when required. The report should contain
recommendations for remedial work when appropriate. For structures incurring no
damage, a simple statement to this effect will be given, unless seismic
instrumentation at the project is activated [47].

Dynamic Analysis for Dams After Earthquakes

For important dams, where great risks exist for human lives and material losses,
dynamic response analysis is worth doing. The United States Bureau of Land
Reclamation (USBR), in its Safety Evaluation of Existing Dams (SEED) Manual,
presents guidelines and procedures to assess the earthquake stability of dams. In
areas of low and infrequent seismic loading, the initial analyses are often conducted
using simplifications and conservatively selected seismic shortcuts data and
assumed properties. If this conservative analysis shows the dam to be safe, no
further work will be required. In areas of more severe and/or frequent seismic
loading, and in the case where an initial simple analysis does not demonstrate the
dam to be safe, more sophisticated analyses are required as explained below.

A. Dynamic Stability (Deformation) Analyses:

Generally, these analyses incorporating the time dependence of the ground
acceleration and the dynamic response will be conducted (i.e., more sophisticated
than Pseudo static) for the various types of dams as follows:
1. Embankment Dams: The initial step will be a simplified analysis utilizing the
NRC (Nuclear Regulatory Commission) response spectrum. Results to be
obtained include the following:
i. The permanent displacements along assumed failure surfaces which extend
through the top one-fourth, the top one-half, and the full height of the
embankment and result from the critical MCE’s (maximum credible
earthquakes) and/or approximate probabilistic earthquakes.
Dam safety and Earthquakes 111

ii. The epicentral distances of magnitude M = 6.5, 7.5, and 8.25 events which
would cause a 3-foot permanent deformation along a failure surface extending
through a critical section of the embankment.
More sophisticated second phase analyses would then be conducted as needed. The
more sophisticated analyses are usually staged (i.e., be progressively more exact)
until either the dam is demonstrated safe or determined to be unsafe which is usually
more difficult.

2. Concrete Gravity Dams: The initial step will be a simplified response

spectrum analysis utilizing the Nuclear Regulatory Commission (NRC)
response spectrum. Local site effects are not considered in the determination
of the spectral amplitudes. Results which are to be obtained include the
following:
i. Peak stresses in critical elements, factors of safety for overturning and sliding,
or analysis by energy methods if the factor of safety is below 1.0 resulting
from the critical MCE’s and approximate probabilistic earthquakes.
ii. The epicentral distances of magnitude M = 6.5, 7.5, and 8.25 events which
cause critical elements to become overstressed, or yield factors of safety of 1.0
for overturning and sliding.

3. Arch and other Concrete Dams: The initial step will be a simplified response
spectrum analysis. Details of the analyses vary from case to case. Results to be
obtained include, as applicable, the ones listed for concrete dams above.
4. Appurtenant Structures: The level of analysis varies from a simple qualitative
assessment to more detailed response spectrum analyses depending on the
importance of the appurtenance to the overall safety of the dam.
B. Liquefaction Analyses: Liquefaction analyses will be conducted for all
foundations and embankments where an initial assessment indicates the
presence of potentially liquefiable materials. The initial analyses would be by
simplified methods. Further, Phase II analyses would be performed to the
extent required.
C. Fault offsets through the Dam and/or Abutments: The effects of fault offsets
would be assessed on a case-by-case basis.
D. Seiche: The effects of seiche arising from ground accelerations (i.e., not
faulting or land sliding) in the reservoir considering oscillations perpendicular
to and parallel to the dam would be investigated. If the earthquakes under
consideration have significant energy content at these periods, then a
simplified modal superposition analysis will be conducted, and the resultant
wave amplitudes estimated. Further second phase analyses would be
accomplished as needed including the effects of overtopping of the dam.
E. Landslides and Fault Displacement Waves: The effects of landslide and
faulting in the reservoir bottom have to be investigated.
112 Nasrat Adamo et al.

10. Rehabilitation of Embankment Dams after Seismic Events

Although many embankment dams have failed during earthquakes, many more have
survived these events but sustained damages of varying magnitudes and remained
operational after rehabilitation. Studying these cases indicates that recent dams
which were designed and constructed with modern techniques have behaved much
better than the older dams. Lessons learned from dam failures, and incidents are
beneficial in avoiding the entire possible shortcomings, while rehabilitation works
carried out indicate possible solutions in similar conditions. The rehabilitation
works of embankment dams after earthquakes are discussed here describing in
general the type of damage(s) and suggesting guidelines for treatment. Each dam,
however, has to be looked at as a unique case and the exact treatment implemented
accordingly.

1. Seismic Evaluation
Proper inspection of dams for any evidence of displacement, bulging, depressions
or undue settlement, tilting, slope instability, cracks, seepage, leakage, erosion and
improper functioning of drains and relief wells, is required immediately after a
seismic event. Any of these conditions, if corrective measures are not taken, can
ultimately lead to failure of the dam. Surface cracks on the crest or near the
embankment or embankment abutment contacts can be an indication of settlement
of the embankment, and if severe enough, a path for leakage can develop along the
contacts. Therefore, these crack locations must be thoroughly examined, and the
following procedures should be adopted for restoration works. In seismic evaluation
of a dam, it has to be ensured that sufficient defensive measures are incorporated by
ensuring good quality control, adequate compaction of materials, foundation and
abutment integrity, ample freeboard, provision of gentle u/s and d/s slope.

2. Repair and Restoration Measures for embankment Dams

Damage to embankment dams varies between, settlements, cracking, increased
seepage and liquefaction of parts of dam or its foundation. Settlement may be
compensated by additional fill or constructing parapet. Other damages may be
treated in prescribed ways. In the following some information is given on such
treatments:
a. General treatment of cracks
Wherever some sort of cracking on dam’s body or separation at its abutments are
visible, they should be investigated in detail, and appropriate remedial works should
be carried out. If the cracks extend some distance along the crest, then test pits are
excavated at suitable intervals to check their depths The trench shall encompass the
whole crack and extends to its full depth and then filled and compacted with the
appropriate material. Further treatment is carried out according to the seriousness
of the crack. An example may be cited from Takami dam in Japan following the
Tokachi-Oki earthquake (M= 8.3) on September 26, 2003. A special inspection after
the earthquake found longitudinal cracks parallel to the dam axis at the crest of the
Dam safety and Earthquakes 113

dam, refer Figure 5. The cracks were up to 160m long in the dam axis direction.
The maximum width of the cracks at the surface was 50mm. An exploratory pit at
the locations of the widest crack was then excavated to measure the depth of the
crack. The results confirmed that the cracks remained within the crest protective
layer with a thickness of about 90cm without reaching the core zone. The excavated
pit was filled in immediately after the survey, and to prevent seepage by rainfall
from expanding the cracks to the core, the entire cracked area was covered with
impervious sheets. One follow-up survey of the crack depth at the crest was carried
out later on as a check. The results confirmed that as in the case of the previous
survey, none of the cracks extended beyond the protective layer. The locations of
the cracks inside the protective layer were repaired by backfilling them with
identical material, refer Figure 6. During the backfilling, density measurements
inside the protective layer were done to control compaction so that the density
would be the same as that before damage [49]. In case the cracks penetrate into the
core, it is suggested to excavate a trench deep enough to expose the top of the
impervious core, the trench to be further dug for 0.5m depths in the impervious
material of the core. The side slopes of the trench are kept at 0.5:1 and bottom width
be kept as 1.0m. The trench in impervious core is to be filled with water and allows
water to seep through. If, at all, any sign of wetness or oozing is found. The trench
is back filled with mixture of impervious soil and 2% to 3% of low-grade bentonite
to be laid and compacted with hand- held pneumatic tempers at 2% of optimum
moisture content and up to a depth of 1.5m. The upper 1.5m layer over core material
is backfilled with semi-pervious type of soil with compaction by pneumatic temper.
The crest is reinstated to its original shape, and riprap is then re-laid as per designed
section. After refilling the trench, 3m c/c holes be drilled and grouted with clay
bentonite slurry at a low pressure if warranted, Figure 7 [50].
b. Treatment for depression and protrusions on d/s face.
If the depressions and/ or protrusions are up to 200mm or less, the downstream
slopes need not to be disturbed. In case the depressions are of larger magnitude, all
material(s) in loose pockets should be removed. The face is to cut to normal to slope
or at about 20; as convenient. Suitable material(s) free from organic material shall
be selected. In choosing material(s) for sand and gravel fill, the material(s) shall be
well grade. The cut surface shall be moistened, and fill material shall be placed in
about 30 cm thick horizontal layers. Each layer is to be compacted by some
pneumatic tamper till its density is about 90 percent of dry density of the material.
The thickness after compaction become about 20cm. Water content should be
uniform throughout the layer to be compacted, and it should be as close as possible
to that content which will result in the maximum densification of the material to be
compacted. In general, this water content will be slightly less than the optimum
water content as determined by Proctor compaction test.
c. Treatment for sloughing
Treatment for sloughing of downstream face is carried out by removing all
undesirable vegetation. A previous blanket or filter material is first laid as buttress
114 Nasrat Adamo et al.

on the face and extended up to the toe filter. Longitudinal cracks observed on the
downstream slope of the dam are to be excavated in the form of a deep trench having
side slope of 0.5:1 and bottom width of 1.0 m keeping the crack in the center. The
trenches are to be backfilled with selected soils with proper rolling and watering.
The existing disturbed pitching on d/s slope is to be re-laid to the designed section
for rain protection.
Dam safety and Earthquakes 115

Figure 5: Takami dam and location of cracks at the Crest [49].

116 Nasrat Adamo et al.

Figure 6: Preparation for Filling the Crack on the Crest of Takami dam [49].

Figure 7: Treatment of longitudinal deep crack at Dam Crest [50].

Dam safety and Earthquakes 117

d. Treatment for longitudinal crack and slip in u/s slope

Where the u/s slope has sloughed down and bulged due to longitudinal slips and
settlement in the earthwork and pitching; the following treatment is to be followed:
To remove first all the pitching work and the damaged earth on the upstream; then
to excavate and remove the earthwork in casing sections in benches of 2 to 3m in
depth and about 4 to 6m in width to facilitate rolling and watering in stages. These
benches are to be excavated step by step, thus clearing the casing zone of all cracks.
Next step shall be to complete the earthwork of the casing zone by taking selected
materials from the borrow pit area, backfilling the excavated portion, after proper
rolling and watering. Then after bringing the u/s slope to the original levels, pitching
is to be duly completed.
e. Treatment of sliding in the upstream protective layer
The case at the Makubetsu Dam, an earth fill dam in Japan, with a height of 26.9m,
completed in 2004, may serve as an example. In this case, the maximum
accelerations of 173.1gal (dam axis direction) and 251.6gal (stream direction) were
recorded in the dam foundation and at the crest, respectively. The epicentral
distance of the dam was 140.6km. The sliding of the protective rock layer on the
earth blanket placed on the left and right bank slopes upstream of the dam body
occurred at two locations; on the left bank, length of the first was 20m, while length
of the second was 23m, these are indicated by numbers 1 and 2 in Figure 8. In
addition to two more locations on the right bank, the first was 80m long and the
second was 30m long, and indicated by the numbers 3 and 4, respectively, in the
same Figure 8. The topmost end of the sliding was directly above the water level
during the earthquake. The location of the maximum sliding is near the spillway on
the right bank side, and the maximum settlement of approximately 90cm was
confirmed by the shape of the slope, and a level difference of about 40cm was
formed at the top edges of the deformation, refer Figure 8. Close up of photos of the
sliding are shown in Figure 9. After the earthquake, the dam management performed
an excavation survey confirming that the sliding was limited to the protective rock
layer (approximately 2m) and did not extend to the interior of the earth`s blanket.
And the part of the protective rock layer, that became thinner because of the sliding,
was thickened by placing more rock materials to temporarily repair the sliding. The
pore water pressure inside the dam body rose slightly at almost all parts of the
upstream side zone, and it had almost peaked at the time of the survey. The values
obtained later by a regular inspection were stabilized, and almost no abrupt change
was found. The seepage was increased slightly due to the earthquake, but it
stabilized later. The dam body displacement in the upstream- downstream direction
near the center of the crest is approximately 2cm on the upstream side, and the
settlement is about 3cm.
118 Nasrat Adamo et al.

Figure 8: Makubetsu Dam layout and cross section showing damaged areas
[49].

Figure 9: (Upper) Layer of Makubetsu Dam and location of slides.

(Lower) views of slides [49].
Dam safety and Earthquakes 119

f. Treatment for seepage

Where heavy seepage is noticed in dams after earthquakes, the complete wet d/s
area is to be cleared of any vegetative growth, particularly the toe drain area. The
seepage water shall have to be channelized, and suitable basins created to measure
the seepage water by providing adequate V-notches at suitable locations. The
seepage is to be properly monitored, particularly, with reference to the reservoir
levels.
In order to assess the final quantity of seepage and also the zones and the area where
the seepage is occurring, whether through the body of the dam or below the
foundation levels; the following measures are suggested: Piezometers may be
installed on the d/s and also in the body of the dam. Monitoring of these would
indicate the nature of seepage, whether through the foundation or the body of the
dam. In case the seepage is mainly in the foundation area, it will be necessary to
provide adequate treatment either by restoring to cut off measures to be suitably
decided after the seepage details are available. Remedial measures for seepage such
as upstream clay blanketing, downstream toe loading, etc. may have to be carried
out in case the seepage is found to be excessive and concentrated at a particular
location at the downstream toe. The seepage collected in the seepage drain at the
toe of the dam should be measured over a V-notch placed at the exit of the dam. In
no case, the seepage should be more than the acceptable limits prescribed by the
design calculations.
g. Strengthening of slopes.
If a section needs strengthening or flattening of slopes, this should be done by
placing coarse pervious material. On the u/s face, rock-fill material may be used, if
possible. On the d/s face, if only limited placement is needed, once again it should
be done by coarse pervious material. However, if substantial placement is required,
the quantity of pervious material required can be reduced by placing a chimney
drain connected to a horizontal drainage filter blanket and rock toe at the end.
Beyond the chimney drain, any clean fill material can be used to give the required
weight. Downstream slope should also be pitched to protect it from rain cuts.
h. Treatment for liquefaction
Where the dam is damaged severely due to liquefaction, the top layer or more as the
case may need, are to be removed from the dam foundation area. The excavated
layer should be replaced by well-compacted semi- pervious material upstream of
the cutoff trench, and well graded compacted pervious material downstream of the
trench. The cutoff trench should be carried down to the rock or hard sub-surface to
provide a positive cutoff. Alternatively, it may be necessary to densify the
foundation material by hammer driven piles on a 3m hexagonal grid through the full
depth of the foundation. This could be done through 15cm minimum diameter pipes
with a flat bottom, which could be extracted, and the hole filled back with pea
gravel. It will be found convenient to use a flat sliding steel shoe at the pipe bottom.
This treatment will have to be done first in a test area and post-treatment N-value
compared with those before the treatment. Spacing of the pipe or its diameter could
120 Nasrat Adamo et al.

be suitably adjusted to achieving the desired N-value. The SPT values should be
aimed at 20 or more, which will be indicative of no further liquefaction. According
to Indian Standards IS: 1893-2002, the recommended N-values are given in Table
7.
Table 7: N-values with depth.
Depth below ground level ≤5m 10 m
Minimum acceptable
15 25
N value, Zone IV and V
Minimum acceptable
10 20
N value, Zone III
Note: For intermediate depths, linear interpolation may be used

Where there are successive slides over the full height of the dam slope, it would be
necessary to remove all the disturbed soil material from upper most slip face to the
required depth. The removed portion can then be rebuilt with a homogeneous low
permeability soil and properly compacted to bring the section to its original
dimensions.

11. Upgrading of Concrete Dams for higher resistance to

Seismic Events
Concrete dams, generally speaking, have fared much better than embankment dams
as far as their resistance to earthquakes. But new studies have shown that they could
fail under the effect of ground shaking and acceleration greater than what they have
experienced. This has necessitated upgrading their designs and performing
additional construction works to shore them up against greater expected events.
Case studies on eight dams that were strengthened to improve their resistance to
seismic shaking are presented in paper authored by Kenneth Hansen, and Larry
Nuss [51] and [52]. All these dams are located in the earthquake prone western
United States and Canada. The studied cases are three concrete arch dams, three
multiple-arch dams, and two slab-and-buttress dams. The basic information on the
dams is contained in Table 8.
These seismic upgrades were based on the knowledge and analytical tools available
at the time. Modifications to these dams in order to improve their seismic
performance were made mainly on the basis of 3-dimensional (3D) linear finite
element analysis (FEA), which produced tensile stresses determined to be in excess
of the predicted tensile strength capacity of the in-place concrete when the structure
is subjected to the Maximum Credible Earthquake (MCE). Only three cases are
described briefly here for illustration while the full scope can be found in the
aforementioned references.
 Dam safety and Earthquakes 121

Table 8: Study cases of dams modifications to strengthen them against seismic events
impacts.
Date Crest Design
Dam Location Height Design
Built Length Dam Type Modification PHGA
feet (m) EQ
Modified feet (m) (g)
Big Bear Valley Multiple Mass concrete
1912 M 8.3
Big Bear Lake, 92 (28) 360 (110) Arch Mass infill between 0.71
1989 at 16 km
CA concrete buttresses
Gibraltar Santa 1920 169 (52) Arch RCC RCC buttres M 7.5
600 (183) 0.68
Barbara, CA 1990 194 (59) buttress downstream at 2 km
Stewart
1930 Full height post- M 7.5
Mountain 212 (65) 583 (178) Arch 0.34
1991 tension cables at 15
Phoenix, AZ
RCC infill
Littlerock 1924 Multiple M 8.0
175 (53) 720 (219) between 0.70
Palmdale, CA 1994 Arch RCC at 2
buttresses
RCC stabilized
Weber 1924 Triple M 6,5
89 (27) 354 (108) sides of central 0.3
Placerville, CA 2002 Arch RCC at 18 km
buttresses
Concrete walls
Stony Gorge 1928 Slab &
139 (42) 868 (265) and struts between M 6.5 0.71
Willows, CA 2009 Buttres
buttresses
Inclined post-
tensioned anchors
for buttresses.
Seymour Falls 1961 Slab &
98 (30) 771 (235) Two concrete M 6.5 0.50
Vancouver, BC 2007 buttres
walls in bays on
either side of
spillway.
Big Tujunga
Concrete buttres M 7.5
Los Angeles, Arch 1.1
down stream at 7 km
CA

1. Bear Valley Dam

The Bear Valley Dam is located in the San Bernardino Mountains about 80 miles
east of Los Angeles. It was the first concrete multiple arch dam to be upgraded to
improve its seismic performance. The dam was completed in 1912 as a 92 foot (28m)
high, 360 foot (110m) long multiple arch dam. In the 1980s, there were concerns
over the structural adequacy of the dam with respect to severe earthquake shaking
as well as overtopping. Two MCEs were considered for the dam rehabilitation. The
most critical shaking was determined to be a M=8.3 earthquake caused by a rupture
on the San Andreas Fault; 10 miles away with a PHGA of 0.45g and a bracketed
duration of 35 seconds (duration between the first and last peaks of 0.5g or greater).
The structural strengthening was accomplished by infilling the arch bays with
conventional mass concrete in order to convert the multiple arch to basically a
gravity dam. The downstream slope was formed at 0.25V:1.0 H except for the
uppermost 47 feet which was vertical (refer Figures 10a and 10b) [52]. The
122 Nasrat Adamo et al.

modification which also included grouting between the old and new concrete to
assure monolithic behavior was completed in 1989. On June 28, 1992, the M=7.4
Landers Earthquake occurred about 28 miles from the strengthened structure. Then
one day later, the closer M=6.6 Big Bear Earthquake occurred about 11 miles from
the dam on an unnamed fault reportedly in response to the rupture on the Landers
Fault. At the Big Bear Lake Civic Center about 2.4 miles from the dam, a strong
motion instrument recorded a peak of 0.57g horizontally and 0.21g vertically during
the Big Bear Earthquake. This recording is about 5 miles closer to the causative
fault than the dam. Thus, the estimated PHGA at the dam was between 0.4 to 0.5g.
A thorough inspection after the earthquake indicated that the dam had not been
damaged. The only damage was a slight displacement of the highway bridge at the
crest of the dam. The bridge has since been removed from the dam and relocated
downstream.

Figure 10: Bear Valley Dam (a) multiple arch dam. (b) Shaded area Shows
concrete infill between arches to improve seismic stability [52].

2. Stewart Mountain Dam

This Dam is located on the Salt River east of Mesa, Arizona. The concrete structure
includes thin arch dam, two thrust blocks, three gravity sections, and one spillway,
Figure 11. The arch has a 212 foot (65m) structural height, an 8 foot (2.4m) crest
thickness, a 34 foot (10m) base thickness, and a 583 foot (178m) crest length.
Construction was completed in March 1930. Concrete placement proceeded without
cooling in 5 foot (1.5m) high blocks. Lack of cleaning horizontal construction joints
before subsequent placements produced weak layers of laitance at every lift surface,
Dam safety and Earthquakes 123

thus resulting in planes with little or no bond. The combination of weak horizontal
planes and vertical contraction joints basically produced independent concrete
blocks. Alkali-aggregate reaction within the concrete has permanently displaced the
arch dam crest 6 inches upstream and 3 inches upward. This reaction has apparently
stopped due to depletion of the alkalis in the cement and is no longer a major
concern. A structural analysis indicated large inertia forces near the crest that are
produced in the 3D linear elastic modal-superposition time-history analysis with
mass-less foundation and (Westergaard Added Mass) using the MCE of Richter
magnitude M=6.75 at 9.3 miles (15 km) occurring on the Sugarloaf Fault. Predicted
horizontal site accelerations are approximately 0.34g upstream/ downstream,
0.258g cross stream, and 0.19g vertical using attenuation curves developed by Seed
and Idriss. The study determined at times, that during the design earthquake, the
upstream inertia force tended to pull portions of the upper arch dam horizontally
apart along the vertical joints. Concrete blocks in these areas may thus be without
much side support from the natural arching action of an arch dam. The arch dam
was believed not to perform dynamically as a monolithic unit as originally intended,
because of the unbonded horizontal lift surfaces. This situation is acceptable for
static loadings because of the horizontal arching action transferring loads along the
arches and 'wedging' the structure in the canyon. The cantilevers still act as
cantilevers because of the transfer of load across the lift surfaces from friction, but
this situation was determined to be not acceptable for dynamic loadings. Large
inertia forces induced in the upstream direction of the dam during an earthquake
were greater than the static reservoir load. It was felt that individual concrete blocks
moving in the upstream direction without arching action would break the frictional
resistance along unbonded lift surfaces causing instability. A progressive failure of
the top central section of the dam was postulated. The dam was modified in 1991
with post-tensioned cables adding the necessary normal force across these
horizontal planes to increase the frictional resistance for stability. The cable
installation consisted of 62 epoxy-coated post-tensioned cables varying from 75 to
240 feet (23 to 73m) in length, inclined between vertical and 8° 40' off vertical. It
is believed that this modification was the first time an arch dam had been post-
tensioned [52].
124 Nasrat Adamo et al.

Figure: 11: (a) Stewart Mountain Dam, (b) Post-tensioned cables installed to
increase frictional resistant along lift lines [52].

3. Big Tujunga Dam

Big Tujunga Dam was a thin-arch concrete dam completed in 1932 on Big Tujunga
Creek as a flood control structure northeast from Los Angeles and is now also used
for water supply. It has a height of 244 feet (74m), a crest length of 830 feet (253m).
The arch is 8-feet (2.4m) thick at the crest and a maximum base thickness of about
74 feet (23m). It is owned by the Los Angeles County Department of Public Works
(LACDPW). The project has required numerous improvement programs, studies,
and remedial actions to address seepage and rock stability issues. In 1975, the owner
conducted a 3D finite element seismic stability study after the 1971 San Fernando
Earthquake had damaged Pacoima Dam about 12 miles away from Big Tujunga
Dam. The seismic time-history record for the MCE used in the study was a modified
recording measured during M=7.5, 1971 event at Pacoima resulting in a PHGA of
1.1g on the Sierra Madre fault. The conclusion from the study was that Big Tujunga
Dam was unable to adequately resist the MCE with a full reservoir. As a result, the
preferred interim alternative was to restrict the reservoir to 77 feet below the
spillway crest, a level determined safe for the MCE by limiting peak tensile stresses
in the dam to below 750lb./in2 as determined to be the maximum dynamic tensile
capacity in 1975 based on core testing. In 1996, MWH Global reanalyzed the
seismic performance of the dam for the MCE and again determined the structure
inadequate. In 1999, LACDPW secured funding to rehabilitate the dam for both
seismic and flood conditions. Of the various concepts to stabilize the dam for a large
seismic event, the preferred alternative was to convert the thin-arch to a thick-arch
Dam safety and Earthquakes 125

dam by adding a conventional concrete overlay against the downstream face; adding
12-feet thickness to the crest and 66-feet thickness to the base, refer Figure 12.

Figure 12: Big Tujunga Dam after modification [52].

The 3D linear-elastic seismic finite element analysis of the modified structure

showed limited isolated areas at the upstream face of the dam of high stress during
a small number of acceleration time history peaks that exceeded the estimated
tensile strength of the concrete and thus not deemed critical. Steel dowels grouted
into the existing dam and protruding into the overlay concrete spaced at 10-foot by
10-foot grid pattern provided additional strength, beyond natural cohesion, between
the new and overlay concrete. Foundation consolidation grouting on 10-foot centers
in each direction reduced the possibility of differential settlement between the
existing and new structures.

12. Conclusion
Earthquakes can have a profound impact on dams’ safety, especially in seismic
active zones. Historical records indicate much dam failure due to earthquake and
even more dams, which suffered varying degrees of damage(s). Factors which have
control over this are first those related to the nature of the event, i.e. the strength of
the earthquake which denotes the amount of energy released and the proximity of
the dam from its epicenter or its location from the fault line, and second those factors
related to the type of dam, its material and method of construction, the design
criteria, and analytical methods used in its design. Most failed dams were tailing
dams or hydraulic fill dams or small earth fill dams which reflect the weight of the
126 Nasrat Adamo et al.

aforementioned factors. Generally speaking, embankment dams are less tolerant to

ground shacking caused by an earthquake than earthquakes. Failure modes or
damage type of embankment dam are due to settlement, cracking or in some cases
liquefaction. Rockfill dam has shown very good performance even when compared
to concrete dams. On the other hand, many concrete dams have sustained various
degrees of the damage(s) and only one gravity dam (barrage) had failed, and that
was a special case of being located on the line of a fault which underwent
considerable vertical and horizontal displacements. As far as damages to concrete
dams, buttress dams have shown less tolerance to earthquakes when compared to
arch and gravity dams. RCC dams which are a recent development in dam design
and construction have shown excellent records although some of them are very large
dams and this is attributed to the nature of construction material, and the up to date
design criteria used. In modern dam designs, engineers have abandoned the obsolete
empirical Pseudo-static method towards a more rational approach known as the
dynamic response method which adopts the expected ground shaking expected from
anticipated earthquakes based on actual seismic records of events in the region
where the dam is located. This has helped considerably to build very large dams in
such a region with more confidence in their safety, whether concrete or earth fill
dams. In all cases, however, the designers of dams are called upon to look into the
many hazards, such as seiches and landslides that may still exist and work to reduce
their impacts on the safety of the dam. Good observation of dam behavior during
and after the earthquake, by visual inspection and examining seismic records helps
considerably in evaluating the dam safety and decide on corrective measures; if
needed. It follows; therefore, that installing seismic instrumentation in dams is very
important and should be provided. Every dam that is built represents a unique case
as far as its reaction towards earthquakes, especially in view of the different ways
that such earthquake can happen; it follows that studying each dam case gives
additional lessons to be added and enrich knowledge in this field and help in design
safer dams.
Dam safety and Earthquakes 127

References
[1] Kinghorn, J. (2019). How Ridiculed Theory of Earth Movement Gained
Ground. Air World Wide Versik Business Web Site entered on 2019- 11- 10
https://www.air-worldwide.com/Blog/How-a-Ridiculed-Theory-of-Earth-s-
Movements-Gained-Ground
[2] Rafferty, J. P. (2019). Seismic Belts. The Encyclopedia Britannica entered on
219-11-09 https://www.britannica.com/science/seismic-belt
[3] Sharmin, J. (2016). Review of Recent Earthquakes in the Light of Plate
Tectonics and Seismic Risk Studies for Bangladesh. LinkedIn Slide Share
Published on April 21, 2016 Slide No45
https://www.slideshare.net/JohanaSharmin/review-of-recent-earthquakes-in-
the-light-of-plate-tectonics-and-seismic-risk-study-for-bangladesh
[4] United State Committee on Large Dams USCOLD (2000a). Observed
Performance of Dams during Earthquakes. Vol II October 2000 PP 15- 20
http://www.ussdams.org/wpcontent/uploads/2016/05/ObservedPerformanceII
_V2.pdf
[5] United State Committee on Large dams USCOLD (2000b) Observed
Performance of Dams during Earthquakes. Vol II October 2000 PP 2- 13
http://www.ussdams.org/wp-
content/uploads/2016/05/ObservedPerformanceII_V2.pdf
[6] Courtney, T., Taylor, C. and Greening, P. (2000). Report from Earthquake
Engineering Field Investigation Team visits to Turkey and Taiwan, following
the recent large earthquakes. 10 February 2000 Water Power and Dam
Construction Magazine
https://www.waterpowermagazine.com/features/featureearthquake-team-
reports-from-turkey-and-taiwan/
[7] Wasson, R. L., Page, R. A., Boore, M. and Yerkes, R. F. (1974). Expected
Earthquakes and Their Ground Motions in the Van Norman Reservoirs Area
Northern San Fernando Valley, California. Geological Survey Circular 691
A, B PP 29- 35 Washington 1974.
https://pubs.usgs.gov/circ/1974/0691a/report.pdf
[8] Sasaki, T., Washita, T. and Yamaguchi, Y. (2007). Calculation of
Hydrodynamic Pressure in Seismic Response Analysis of Gates. Dam
Structures Research Team, Hydraulic Engineering Research Group, Public
Works Research Institute, Tsukuba-shi, Ibaraki-ken, 305-8516 Japan
https://www.pwri.go.jp/eng/ujnr/joint/39/paper/62iwashita.pdf
[9] United State Committee on Large Dams USCOLD (2014). Observed
Performance of Dams during Earthquakes. Vol III Page 3 and PP 43- 44
February 2014
file:///C:/Users/HP/Documents/A%20Dam%20Safety%20Book/Dams%20an
d%20Eartquakes/2014%20Observed_Performance_of_Dams_During_Eartqu
akes%20vol%203%20%202014.pdf
128 Nasrat Adamo et al.

[10] United State Committee on Large dams USCOLD (2014). Observed

Performance of Dams during Earthquakes. Vol III Page 5 February 2014
file:///C:/Users/HP/Documents/A%20Dam%20Safety%20Book/Dams%20an
d%20Eartquakes/2014%20Observed_Performance_of_Dams_During_Eartqu
akes%20vol%203%20%202014.pdf
[11] Hinks, J. (2015). Dams and earthquakes. SECED 2015 Conference:
Earthquake Risk and Engineering towards a Resilient World 9-10 July 2015,
Cambridge UK
file:///C:/Users/HP/Documents/A%20Dam%20Safety%20Book/Dams%20an
d%20Eartquakes/Dams%20and%20EArthquakes%20Hink.pdf
[12] Bosshard, P. (2009). The Dam that Shook the World. International Rivers
blog Page entered on 2019-11-25.
https://www.internationalrivers.org/blogs/227/the-dam-that-shook-the-earth
[13] United State Committee on Large dams USCOLD (2014). Observed
Performance of Dams during Earthquakes. Vol III Page 6 February 2014
file:///C:/Users/HP/Documents/A%20Dam%20Safety%20Book/Dams%20an
d%20Eartquakes/2014%20Observed_Performance_of_Dams_During_Eartqu
akes%20vol%203%20%202014.pdf
[14] EERI Special Earthquake Report (2010). Learning from Mw 8.8 Chile’s
Earthquake of February 27, 2010. Page 5, June 2010
https://www.eeri.org/site/images/eeri_newsletter/2010_pdf/Chile10_insert.pd
f
[15] Wieland, M. (2010). Features of Seismic hazards in large dam projects and
strong motion monitoring of Large Dams. Frontiers of Architecture and Civil
Engineering in China 4(1):56-64 DOI: 10.1007/s11709-010-0005-
https://www.researchgate.net/publication/226446090_Features_of_seismic_h
azard_in_large_dam_projects_and_strong_motion_monitoring_of_large_dam
s
[16] United State Committee on Large dams USCOLD (2000c) Observed
Performance of Dams during Earthquakes. Vol II October 2000 PP10- 11.
http://www.ussdams.org/wp-
content/uploads/2016/05/ObservedPerformanceII_V2.pdf
[17] Hansen, K.D. and Nuss, L.K. (2011). Lessons learned from the earthquake
performance of concrete dams. Water Power and Dam Construction
Magazine 27 September 2011
https://www.waterpowermagazine.com/features/featurelessons-learned-from-
the-earthquake-performance-of-concrete-dams/
[18] Nuss, L.K., Hansen, K.D., Matsumoto, N. and Lyvers, G.M. (2012) and (2017).
Shaken but not stirred-Earthquake Performance of Concrete Dams.
Presentation from USSD conference New Orleans April 2012 (Session 1), and
included in the Seismic Analysis of Concrete Dams Workshop (2017 USSD
annual conference, April 6-7-2017), http://www.ussdams.org/wp-
content/uploads/2016/05/Session1_Case-Histories-and-Seismic-PFM.pdf
Dam safety and Earthquakes 129

[19] Wikipedia (2019). The 1967 Koynanagar earthquake. Website page entered
on 2020-04-01
https://en.wikipedia.org/wiki/1967_Koynanagar_earthquake
[20] Matsumoto, N., Sasaki, T. and Omahci, T. (2011). The 2011 Tohokou
Earthquakes and Dams. Proceeding of the ICOLD 89th Annual Meeting I
Lucerne June 1, 2011 http://www.jcold.or.jp/bk_e/activity/Tohokuswiss.pdf
[21] Salmon, J.W., Nuss, L. and Dominic, S. (2011) Seismic Induced Loads on
Spillway Gates- Phase 1 Literature Review. USBR Report DSO- 11- 06 PP
3-4Technical Service Center
file:///C:/Users/HP/Documents/A%20Dam%20Safety%20Book/Dams%20an
d%20Eartquakes/DSO-11-06.pdf
[22] National Research Council (1985). Safety of Dams-Flood and Earthquake
criteria. Committee on Safety Criteria for Dams, pp. 32-34 National Academy
Pressfile:///C:/Users/HP/Documents/A%20Dam%20Safety%20Book/Dams%
20and%20Eartquakes/Dam%20Safety%20Floods%20and%20Earthquakes.p
df
[23] Wieland, M. (2013). Seismic Design of Components. Water Power and Dam
Construction 22 April 2013
https://www.waterpowermagazine.com/features/featureseismic-design-of-
major-components/
[24] Wieland, M. (2018). Application of Pseudo-static Analysis in Seismic Design
and Safety Evaluation of Embankment Dams, pp. 3-4, Earthquake
Engineering Conference, Thessaloniki 18-22 June 2018.
http://papers.16ecee.org/files/Paper%2016ecee%20Wieland%20ID%20136%
20pseudostatic%20analysis%202018%20revised.pdf
[25] Wieland, M. (2012a). Seismic Design and Performance Criteria for Large
Storage Dams. Proceeding of the Fifteenth World Conference on Earthquakes
Lisbon 2012 Vol 5 PP. 3677-3686 Paragraph 5 Lisbon 2012
https://www.iitk.ac.in/nicee/wcee/article/WCEE2012_0638.pdf
[26] Wieland, M. (2011). Presentation of Seismic Design and Performance Criteria
for Large Dams. Poyry Switzerland Ltd Zurich Switzerland
https://pdfs.semanticscholar.org/a491/9d64b662c3a89dcaaec7400f2b400863
9c1c.pdf
[27] Wieland, M. (2012b). Dam Safety and Earthquakes. Waterpower and Dam
Construction September 2012
https://www.waterpowermagazine.com/features/featuredam-safety-and-
earthquakes
[28] Hinks, J. (2015). Dams and Earthquakes. Proceedings of the 2015
Conference of the Society of Earthquake and Civil Engineering Dynamics:
Earthquake Risk and Engineering towards Resilient World 9-10 July, 2015,
Homerton College, Cambridge.
file:///C:/Users/HP/Documents/A%20Dam%20Safety%20Book/Dams%20an
d%20Eartquakes/Dams%20and%20EArthquakes%20Hink.pdf
130 Nasrat Adamo et al.

[29] Lyu, Z., Chai, J., Xu, Z., Qin, Y. and Cao, J. (2019) A comprehensive
Review on Reasons for Tailing Dams Failures based on Case History.
Hindawi Advances in Civil Engineering Journal, Vol. 2019, Para 2.4,
https://www.hindawi.com/journals/ace/2019/4159306/ or (Pdf copy) PP 7- 9
http://downloads.hindawi.com/journals/ace/2019/4159306.pdf
[30] Sing, R., Roy, D. and Jain, S. (2015). Investigation of Liquefaction Failures
in Earthen Dams During Bhuj Earthquake. Indian Institute of Technology
Department of Civil engineering 2015.
http://www.iitk.ac.in/nicee/RP/2005_Earthen_Dams_CBIP.pdf
[31] Adlier, K. and Sharp, M. (2004). Dynamic Behavior of Embankment on
Liquefiable Foundation subject to Moderate Earthquake loading 13th World
Conference on Earthquake Engineering Vancouver, B.C., Canada August 1-6,
2004 Paper No. 1025 https://www.iitk.ac.in/nicee/wcee/article/13_1025.pdf
[32] Wieland, M. (2008). Large Dams the First Structures Designed Systematically
Against Earthquakes. Paragraph 5, Proceedings of the 14th world conference
on Earthquakes Engineering, October 12-17, 2008, Beijing, China
http://www.iitk.ac.in/nicee/wcee/article/14_S13-049.PDF
[33] Lakshmanan, C.T. (2012). Landslides. Slide No. 18 of presentation of SRM
School of Architecture SRM University Published on May 25 March, 2012
https://www.slideshare.net/ctlachu/14-landslide?next_slideshow=1.
[34] Ghumare, A. (2014). Landslide Hazard Zonation Mapping Using RS and GIS.
Slide No. 8 Dar Al-Handasa (Share and Partners), October 21, 2014
https://www.slideshare.net/adityaghumare/landslide-hazard-zonation-
mapping-using-rs-and-gis?next_slideshow=3
[35] Petley, D. (2013). Landslides and Large Dams. Presentation given in the
International Conference on Vajont incidence (1963) held in Padua 8- 10,
October 2013.
https://www.slideshare.net/Dr_Dave/landslides-and-large-dams
[36] Petley, D. (2013). Landslides and Large Dams. The full paper on Vajont
2013 conference.
https://blogs.agu.org/landslideblog/2013/11/20/landslides-and-large-dams/
https://blogs.agu.org/landslideblog/2013/11/20/landslides-and-large-dams/
[37] Petley, D. (2013). Global Losses from Landslides Associated with Dams and
Reservoirs. Durham University - Department of Geography-Durham, UK,
DOI:10.4408/IJEGE.2013-06.B-05 Italian Journal of Engineering Geology
and Environment - Book Series (6) Page 69 October 10, 2013 Sapienza
Università Publishing.
http://www.ijege.uniroma1.it/rivista/international-conference-on-vajont-
1963-2013-thoughts-and-analyses-after-50-years-since-the-catastrophic-
landslide/keynotes/global-losses-from-landslides-associated-with-dams-and-
reservoirs/ijege-13_bs-petley.pdf
Dam safety and Earthquakes 131

[38] Gohari, M. and Avarideh, A. (2019). The Effect of Landslide at Dam

Reservoirs on Seismic Behavior of Earth Dams. 4th International Conference
on Sustainable Architecture and Urbanism. Paper retrieved from Research
Gate on 2019-12- 07
https://www.researchgate.net/publication/324573466_The_Effect_of_Landsli
de_at_Dam_Reservoirs_on_Seismic_Behavior_of_Earth_Dams
[39] Barjasteh, A. (2019). Seismicity Triggered Landslide Potential in the
Reservoir of Gotvand Dam, Iran. 8th International conference on seismology
and earthquake engineering ,11-13, Tehran, Iran.
http://www.iiees.ac.ir/see8/SEE8-GH00910264.pdf
[40] Wieland, M. (2012c). Seismic Design and Performance Criteria for Large
Storage Dams. Proceeding of the Fifteenth World Conference on Earthquakes
Lisbon, 2012, Vol 5, pp. 3677-3686. Paragraph 3, Lisbon 2012
https://www.iitk.ac.in/nicee/wcee/article/WCEE2012_0638.pdf
[41] Schuster, R. (2006). Interaction of Dams and Landslides-Case Studies and
Mitigation. USGS Professional Paper 1723, Reston,Virginia: 2006
https://pubs.usgs.gov/pp/2006/1723/pdf/P1723_508.pdf
[42] Layton, J. (2020). Can humans start an earthquake?
https://science.howstuffworks.com/environmental/green-science/humans-
start-earthquake2.htm
[43] Pradeep, T. and Liu, S. (2011). Reservoir-induced seismicity in the
Danjiangkou Reservoir: a quantitative analysis Geophysical Journal
International
https://www.academia.edu/26629081/Reservoir-
induced_seismicity_in_the_Danjiangkou_Reservoir_a_quantitative_analysis
[44] Pedro Simão Sêco e Pinto (2010). Understanding Seismic Embankment Dam
Behavior through Case Histories Fifth International Conference on Recent
Advances in Geotechnical Earthquake Engineering and Soil Dynamics,
Missouri University of Science and Technology Scholars' Mine
https://scholarsmine.mst.edu/cgi/viewcontent.cgi?article=2937&context=icra
geesd;Understanding ]
[45] Evangelidis, C. (2019). Seismic monitoring of Large Dams. Alliance for
Disaster Risk Reduction 12th August 2019
http://alter-project.eu/wp-
content/uploads/2018/12/ALTER_seismic_monitoring.pdf
[46] Pandya, V. (2013). The Importance of Seismic Instrumentation in Dam
Monitoring. Journal of Information, Knowledge and Research in Civil
Engineering. ISSN: ISSN 0975–6744 Period: Nov.12 To Oct. 13, Volume
2, Issue 2 pp. 99-103.
http://www.ejournal.aessangli.in/ASEEJournals/CIVIL18.pdf
The article may be downloaded directly from
http://www.ejournal.aessangli.in/CivilEngineering.php CIVIL18
 132 Nasrat Adamo et al.

[47] United State Army Corps of Engineers (2017). Engineering and Design-
Reporting Earthquake Effects. ER 1110-2- 1802 28 July 2017
https://www.publications.usace.army.mil/Portals/76/Publications/EngineerRe
gulations/ER_1110-2-1802.pdf
[48] United State Bureau of Reclamation (1995). Safety Evaluation of Existing
Dams. Manual for the Safety Evaluation of Embankment and Concrete Dams
Denver Colorado 1995 Reprint 2000 pp. 75-77.
https://www.usbr.gov/tsc/techreferences/mands/mands-pdfs/SEED.pdf
[49] Nagayama, I., Yamaguchi, Y., Sasaki, T., Nakamura, A., Kawasaki, H., and
Hirayama, D. (2003) and (2019.) Damage to Dams due to Three Large
Earthquakes Occurred in 2003, in Japan. May 26, 2003.
file:///C:/Users/HP/Documents/A%20Dam%20Safety%20Book/Damage%20
to%20Dams%20due%20to%20Three%20Large%20Earthquakes%20Occurre
d%20in%202003,%20in%20Japan.pdf
Or Published again on June 6, 2019 on
https://pdfslide.net/documents/damage-to-dams-due-to-three-large-
earthquakes-occurred-in-damage-to-dams.html
[50] Paul, D. K. (2004). Assessment of Damage and Rehabilitation of Small and
Medium Earth Dams. 13th World Conference on Earthquake Engineering
Vancouver, B.C., Canada August 1-6, 2004, Paper No. 2160.
http://www.iitk.ac.in/nicee/wcee/article/13_2160.pdf
[51] Hansen, K. D. and Nuss, L.K. (2011). Lessons Learned from the Earthquake
Performance of Concrete Dams. September 27, 2011.
https://www.waterpowermagazine.com/features/featurelessons-learned-from-
the-earthquake-performance-of-concrete-dams/
[52] Hansen, K. D. and Nuss, L. K. (2013). Seismic Upgrades for Concrete Dams
– Then and Now, Final July 12, 2013
https://www.academia.edu/6473072/Seismic_Upgrades_for_Concrete_Dams
_-_Then_and_Now_Final_Hansen_and_Nuss_July_12_2013

View publication stats