Anda di halaman 1dari 153

ANCOLD Guidelines

for Design of Dams and

Appurtenant Structures
for Earthquake
Draft Version for publishing on website
March 2017



To the maximum permitted by law, each of, ANCOLD Incorporated and its
Members, the Convenor and Members of the Working Group which developed

these Guidelines, and the Independent Reviewers of these Guidelines exclude all
liability to any person arising directly or indirectly from that person using this
publication or any information or material contained within it. Any person acting on
anything contained in, or omitted from, these Guidelines accepts all risks and
responsibilities for losses, damages, costs and other consequences resulting
directly or indirectly from such use and should seek appropriate professional
advice prior to acting on anything contained in the Guidelines.

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 1
The committee members were;

Mr Steve O’Brien (Consultant, VIC - Convenor)
Mr Ian Landon Jones (Consultant, NSW)
Emeritus Professor Robin Fell (Consultant, NSW)
Mr Brian Cooper (Consultant, NSW)
Mr Jon Williams (Consultant, QLD)
Mr Peter Allan (Regulator, QLD)
Mr Graeme Bell (Consultant, NSW)
Mr David Ryan (Consultant, QLD)
Mr Gary Gibson (Consultant, VIC)
Dr Gavan Hunter (Consultant, VIC)
Mr Andrew Reynolds (Owner, NSW)
Mr David Brett (Consultant, NSW)
Dr Paul Somerville (Consultant, NSW/USA)
Mr Bundala Kendaragama (Consultant, VIC)

Mr Dan Forster (Consultant, NZ)
Dr Nihal Vitharana (Consultant, NSW)
The following members formed the expert review panel:


Additional inputs and review comments were provided by:



Technical Editor


ANCOLD Executive Liaison


ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 2
Table of Contents

1 INTRODUCTION ......................................................................................................................................... 6
2.1 Earthquakes and Their Characteristics .................................................................................................... 8
2.2 Terminology .......................................................................................................................................... 8
2.3 General Principles for Selection of Design Ground Motion and Analysis Method ................................... 8
2.4 Description of a Probabilistic Seismic Hazard Assessment ..................................................................... 9
2.5 Requirements of a Seismic Hazard Assessment ...................................................................................... 9
2.6 Selection of Design Seismic Ground Motion - Deterministic Analysis Approach .................................. 11
2.7 Selection of Design Seismic Ground Motion - Risk Based Analysis Approach ..................................... 12
2.8 Modelling Vertical Ground Motions ..................................................................................................... 12
2.9 Selection of Response Spectra and Time-History Accelerograms.......................................................... 12
2.10 Earthquake Aftershocks ....................................................................................................................... 12

2.11 Seismic Ground Motions from Earthquakes Induced by the Reservoir .................................................. 12
3.1 Effect of Earthquakes on Embankment Dams ....................................................................................... 13
3.2 Defensive Design Principles for Embankment Dams ............................................................................ 13
3.3 Seismic Deformation Analysis of Embankment Dams .......................................................................... 15
3.4 Assessment of the Effects of Liquefaction in Embankment Dams and Their Foundations ..................... 16
3.4.1 Overall Approach ......................................................................................................................... 16
3.4.2 Definitions and the Mechanics of Liquefaction ............................................................................. 17

3.4.3 Methods for Identifying Soils which are Susceptible to Liquefaction ............................................ 17
3.4.4 Methods for Assessing Whether Liquefaction may occur .............................................................. 17
Assessing the Strength of Liquefied Soils in the Embankment and Foundation.............................. 18

3.4.6 Methods for Assessing the Post-Earthquake Strength of Non-Liquefied Soils in the Embankment
and Foundation ............................................................................................................................................ 18
3.4.7 Site Investigations Requirements and Development of Geotechnical Model of the Foundation ...... 18
3.4.8 Liquefaction Analysis ................................................................................................................... 18
3.4.9 Assessment of the Likelihood and the Effects of Cracking of Embankment Dams Induced by
Seismic Ground Motions ............................................................................................................................. 18
3.5 Methods for Upgrading Embankment Dams for Seismic Ground Motions ............................................ 19
3.5.1 General Approach ........................................................................................................................ 19
3.5.2 Embankment Dams not Subject to Liquefaction ............................................................................ 19
3.5.3 Embankment Dams Subject to Liquefaction.................................................................................. 19
3.6 Flood Retarding Basins ........................................................................................................................ 20
4 ASSESSMENT OF CONCRETE DAMS FOR SEISMIC GROUND MOTIONS ........................................ 21

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 3
Table of Contents

4.1 Effect of Earthquakes on Concrete Dams.............................................................................................. 21

4.2 Defensive Design Principles for Concrete Dams ................................................................................... 21
4.3 General Principles of Analysis for Seismic Ground Motions ................................................................. 22
4.4 Seismic Structural Analysis of Concrete Dams ..................................................................................... 22
4.4.1 Material Properties ....................................................................................................................... 22
4.4.2 Loads ........................................................................................................................................... 23
4.4.3 Methods Available and When to Use Them .................................................................................. 24
4.4.4 Analysis in the Frequency Domain ............................................................................................... 26
4.4.5 Analysis in the Time Domain ....................................................................................................... 26
4.5 Approach to Analyses and Acceptance Criteria .................................................................................... 26
4.5.1 During the Earthquake .................................................................................................................. 26
4.5.2 Post-Earthquake ........................................................................................................................... 26

4.5.3 General ........................................................................................................................................ 27
5 ASSESSMENT OF TAILINGS DAMS FOR SEISMIC GROUND MOTIONS ........................................... 28
5.1 Some General Principles ...................................................................................................................... 28
5.2 The Effects of Earthquakes on Tailings Dams ...................................................................................... 28
5.3 Defensive Design Principles for Tailings Dams .................................................................................... 28
5.4 Design Seismic Ground Motions and Analysis Method ........................................................................ 29
5.5 Seismic Deformation Analysis of Tailings Dams not Subject to Liquefaction ....................................... 30
5.6 Assessment of the Effects of Liquefaction in Tailings Dams and Their Foundations ............................. 30

5.7 Methods for Upgrading Tailings Dams for Seismic Loads .................................................................... 30
The Effects of Earthquakes on Appurtenant Structures ......................................................................... 31

6.2 Intake Towers ...................................................................................................................................... 31
6.2.1 General ........................................................................................................................................ 31
6.2.2 Defensive Design Principles ......................................................................................................... 32
6.2.3 Performance Requirements ........................................................................................................... 32
6.2.4 Analysis Procedures ..................................................................................................................... 32
6.3 Spillways ............................................................................................................................................. 33
6.3.1 General ........................................................................................................................................ 33
6.3.2 Defensive Design Principles ......................................................................................................... 34
6.3.3 Performance Requirements ........................................................................................................... 34
6.3.4 Analysis Procedures ..................................................................................................................... 34
6.4 Spillway Gates ..................................................................................................................................... 35

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 4
Table of Contents

6.4.1 General ........................................................................................................................................ 35

6.4.2 Defensive Design Principles ......................................................................................................... 35
6.4.3 Performance Requirements ........................................................................................................... 35
6.4.4 Analysis Procedures ..................................................................................................................... 35
6.5 Outlet Works - Water Conduits, Gates and Valves ................................................................................ 36
6.5.1 General ........................................................................................................................................ 36
6.5.2 Defensive Design Principles ......................................................................................................... 36
6.5.3 Performance Requirements ........................................................................................................... 36
6.5.4 Analysis Procedures ..................................................................................................................... 37
6.6 Retaining Walls ................................................................................................................................... 37
6.6.1 General ........................................................................................................................................ 37
6.6.2 Defensive Design Principles ......................................................................................................... 37

6.6.3 Performance Requirements ........................................................................................................... 37
6.6.4 Analysis Procedures ..................................................................................................................... 37
6.7 Parapet Walls ....................................................................................................................................... 38
6.7.1 General ........................................................................................................................................ 38
6.7.2 Defensive Design Principles ......................................................................................................... 38
6.7.3 Performance Requirements ........................................................................................................... 38
6.7.4 Analysis Procedures ..................................................................................................................... 38
6.8 Mechanical and Electrical Equipment ................................................................................................... 38

6.8.1 General ........................................................................................................................................ 38

6.8.2 Defensive Design Principles ......................................................................................................... 38
Performance Requirements ........................................................................................................... 39

6.8.4 Analysis Procedures ..................................................................................................................... 39
6.9 Access Roads and Bridges .................................................................................................................... 39
6.9.1 General ........................................................................................................................................ 39
6.9.2 Defensive Design Principles ......................................................................................................... 39
6.9.3 Performance Requirements ........................................................................................................... 39
6.9.4 Analysis Procedures (Bridges) ...................................................................................................... 40
6.10 Reservoir Rim Instability ..................................................................................................................... 40

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 5
1.0 Introduction

1 INTRODUCTION The Guidelines provide detailed descriptions of

The purpose of these Guidelines is to provide these and guidance on the requirements of
Owners, Regulators, Consultants, and others seismic hazard assessments.
involved in the management of existing and
new dams and appurtenant structures guidance There have also been advances in the methods
on the selection of seismic ground motions available for analysis and design including:
resulting from earthquakes, and analysis and Embankment dams:
assessment procedures. The Guideline covers · Assessment of the potential for liquefaction
water supply dams, tailings dams and retarding of dams and their foundations, and post-
basins (levees are not covered in this earthquake liquefied strengths.
Guideline). · Numerical modelling of deformations
The Guidelines are not: during seismic ground motions and post-
· A textbook on earthquakes and design for
seismic ground motions. Concrete Dams
· A design code or Australian Standard. · Numerical modelling of the dams for
Hence the mandatory “shall” or “ensure” seismic ground motions including
modelling of deformations and

has been avoided in the text.
These Guidelines have been written for the · Improved understanding of the importance
guidance of experienced practitioners. There is of identifying and modelling kinematically
a need for the users of the Guideline to apply viable mechanisms for sliding in the
their own experience and judgement in the foundation.
application of this Guideline. Therefore, it is
strongly recommended that personnel who Appurtenant Structures
undertake, or at least closely supervise, the · Numerical modelling of the structures for
assessment procedures outlined within this seismic ground motions including
Guideline are highly experienced in field of the modelling of deformations and

assessment being undertaken.

There have been significant advances in the Importantly, there is a greater understanding
understanding of earthquakes in Australia since and expertise within the Australian dam
ANCOLD (1998) “Guidelines for Design of industry than in 1998.

Dams for Earthquake” was published. These

Given the extensive literature relating to
seismic hazard assessment, and seismic
· A recognition that there are active faults in analysis and design, these Guidelines retain a
some areas, and other neotectonic faults large body of information to assist users. These
which have been active in the current are in the Commentary. Users should read the
regional stress regime and may contribute Commentary along with the Guideline Sections
to the seismic hazard. so they understand the details and background
· Refinement of earthquake source models to the Guideline.
for modelling seismic hazard. Since the issue of ANCOLD (1998) a number
· Refinement of ground motion models, and of other ANCOLD Guidelines have also been
development of models for Australian updated. Where appropriate, references have
conditions. been made to these other Guidelines rather than
· Refinement in the modelling of uncertainty repeating sections of them. If inconsistencies
in the seismic hazard. occur then this Guideline supersedes older

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 6
1.0 Introduction

The Guideline encourages the use of risk based

methods for assessing existing and design of
new dams. This is because of the uncertainty in
the seismic hazard, and the fact that the
response of dams and appurtenant structures to
seismic loads may change significantly with
the seismic ground motion. Therefore, a dam or
appurtenant structure that is assessed using a
single design ground motion only may not
necessarily satisfy ANCOLD (2003, 2017
(pending)) Risk Management Guidelines.
However, it is recognised that many Owners
prefer to use a deterministic approach to
assessment of design seismic ground motions
so both approaches are covered in the
For the purposes of defining roles and

responsibilities the following terminology has
been adopted throughout this Guideline:
· Owner – the entity responsible for the asset,
including operation and maintenance and
holds a duty of care to the community that
the asset risks are as low as reasonably
practicable .
· Regulator – the entity responsible for
regulation of dams within each state or

· Consultant – the entity responsible for
provision of assessment and design services
referred within the Guideline. This role

may be performed by the Owner,

educational or government institution, or by
a professional consulting entity suitably
qualified to deliver the services.
AS1170 (Structural Design Actions - Part 4:
Earthquake actions in Australia) specifically
states that dams are not within the scope of the
standard and therefore AS1170 should not be
used for dams and appurtenant structures.

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 7
2.0 Assessment Of Design Seismic Ground Motions And Analysis Method

2 ASSESSMENT OF DESIGN SEISMIC the presently known or presumed tectonic

Maximum Credible Earthquake Ground
2.1 Earthquakes and Their Characteristics Motion – The most severe ground motion
Users of this Guideline should read the affecting a dam site due to an MCE.
Commentary on earthquakes and their Active Faults – a fault, reasonably identified
characteristics generally and Australian and located, known to have produced historical
earthquakes in particular. A good understanding earthquakes or showing evidence of movements
of these is important to using the Guideline. in Holocene time (i.e. in the last 11,000 years ),
2.2 Terminology large faults which have moved in Latest
Pleistocene time (i.e. between 11,000 and
The following terminology is based on 35,000 years ago).
ANCOLD (1998), and ICOLD Bulletin 148 -
Selecting Seismic Parameters for Large Dams Neotectonic fault - a fault that has hosted
(ICOLD, 2016a). displacement under conditions imposed in the
current crustal stress regime, and hence may
Operating Basis Earthquake (OBE) – the OBE move again in the future.

is that level of ground motion at the dam site for
which only minor damage is acceptable. The 2.3 General Principles for Selection of
dam, appurtenant structures and equipment Design Ground Motion and Analysis
should remain functional and damage from the Method
occurrence of earthquake shaking not exceeding There are two ways of selecting the design
the OBE should be easily repairable. seismic ground motion and analysing the
Safety Evaluation Earthquake (SEE) – the structures:
SEE is the maximum level of ground motion for (a) Deterministic Analysis – This requires the
which the dam should be designed or analysed. selection of an OBE and SEE, taking into
Damage can be accepted but there should be no consideration the Consequence Category of

uncontrolled release of water from the reservoir the dam, and the implications of failure of
or tailings from tailings dams. appurtenant structures. The dam and
The term Safety Evaluation Earthquake appurtenant structures are analysed for these
replaces the term Maximum Design Earthquake ground motions. For the SEE it is usual to

(MDE) in ANCOLD (1998, 2012a, 2013). require a factor of safety to be applied to

estimated stresses and deformations within
The OBE and SEE are expressed in terms the dam and its foundations to give a low
relevant to the design of the dam or appurtenant likelihood of the dam failing given the SEE
structure. It may be in terms of peak ground loading.
acceleration (PGA), peak ground velocity (b) Risk Based Analysis– which assesses the
(PGV), or response spectra, which may be effects on the dam and its foundations of a
accompanied by ground motion time-histories range of seismic loads up to and greater
or by proxy information such as moment than the SEE. The design requirement is to
magnitude Mw representing ground motion satisfy ANCOLD (2003, 2016) tolerable
duration. risk criteria. The steps in the process are
detailed in Section C2.3.
Maximum Credible Earthquake (MCE) – the
MCE is the largest reasonably conceivable Preferred method – Based on consideration of
earthquake magnitude that is considered all the factors ANCOLD prefers the risk based
possible along a recognised fault or within a approach to the deterministic approach for High
geographically defined tectonic province, under and Extreme Consequence Category dams. For

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 8
2.0 Assessment Of Design Seismic Ground Motions And Analysis Method

Significant and Low Consequence Category The seismic hazard assessment should be
dams, deterministic approaches will usually be carried out by Seismologists experienced in
adequate. seismic conditions in Australia, which are
different in some regards to what occur
The decision as to whether a risk based or elsewhere.
deterministic approach is adopted is for the
When requesting a seismic hazard assessment
Owner in consultation with the Regulator.
the Owner or Consultant should provide site
The Consequence Category of the dam is specific details to the Seismologist that may
assessed in accordance with ANCOLD (2012). include latitude and longitude for the site,
natural frequency of the structures being
2.4 Description of a Probabilistic Seismic assessed, specific return periods to be included
Hazard Assessment in the seismic hazard assessment, minimum
A Probabilistic Seismic Hazard Assessment magnitudes, geological information for the site,
(PSHA) is an evaluation of the ground Vs30 data, etc.
motion level that will be exceeded at a
specified frequency or annual probability. The seismic hazard assessment for High and
PSHA involves the following steps: Extreme Consequence Category dams should
include the following features:

· Identify all earthquake sources capable
1. Distributed earthquake source models.
of generating damaging ground motions
Analyses should use all viable source
at the site, and characterise their
models. The weighting between these
location, geometry and sense of slip;
should be determined by the Seismologist in
· Characterise the rates at which
consultation with the Owner and
earthquakes of various magnitudes are
expected to occur on each source;
2. Fault sources. Active and neotectonic faults
· Characterise the distribution of source- which could significantly contribute to the
to-site distances associated with ground motion for the dam should be
potential earthquakes; identified, and be accounted for in the

· Predict the resulting ground motion seismic hazard assessment.

level for each earthquake scenario using 3. Faults in the dam foundation. Information
ground motion prediction equations; on any known active or neotectonic faults
· Combine uncertainties in earthquake which have the capability to cause

size, location and ground motion level displacement in the foundation of the dam,
using the total probability theorem appurtenant structures, or reservoir rim
within a qualified computer program. should be provided.
4. Ground motion prediction models.
Section C2.4 gives a detailed description of the Consideration of multiple alternative ground
PSHA process. motion prediction equations (GMPE’s) to
address epistemic uncertainty. These may
2.5 Requirements of a Seismic Hazard
Assessment vary for cratonic and non-cratonic areas.
For non-cratonic areas at least one of the
A PSHA is required for both deterministic and coefficients should be Next Generation
risk based approaches. Attenuation (NGA), along with one or more
Where there are active fault(s) in the vicinity of of the models developed for Australia.
the dam site the deterministic method for Weightings applied to the models should be
Extreme Consequence Category dams also determined by the Seismologist in
requires assessment of the ground motion at the consultation with the Owner and
dam site from the MCE on the active fault(s). Consultant.

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 9
2.0 Assessment Of Design Seismic Ground Motions And Analysis Method

5. Shear wave velocity of the dam foundation fractile), 85th fractile and 95th fractile
bedrock. Shear wave velocity of the percentage spectral response spectra plots.
foundation rock should be determined and 10. De-aggregation plots and selection of time-
included in the ground motion evaluation. history motions. The report should include
The shear wave velocity may vary from plots showing de-aggregation of the hazard
valley to abutment sections so the seismic by earthquake magnitude and distance, and
hazard may be different in the valley and based on these contributions provide
the abutments. guidance on how to select time-history
6. Effects of soil overlying bedrock. If the dam motions (accelerograms) suited for the dam
is founded on sediments or deep residual and the appurtenant works.
soils and completely weathered rock 11. Reservoir-induced seismicity. For new
overlying the bedrock which are expected to dams, guidance on whether reservoir-
have significant nonlinear response, the induced seismicity need be considered, and
PSHA ground motions should be specified if so, the resultant seismic loading.
at the surface of bedrock beneath the
sediments or weathered rock at a depth in For Significant Consequence Category dams,
the profile having a suitable shear wave Owners may opt with the advice of their
velocity that is identified jointly with the Seismologist and Consultant to not require

soil response Consultant, to provide input items (2), (3), and (11); only one GMPE (4),
into a separate study of nonlinear soil and use estimated shear wave velocities (6); and
response that would not be part of the provide mean estimates without modelling
uncertainty (9). However if that study shows
seismic hazard assessment. The bedrock
ground motions may be amplified or de- that the seismic hazard is critical to the
amplified by the overlying soil depending assessment of the risks posed by the dam the
on the soil depth and properties and the more complete assessment of the seismic
level of the bedrock seismic ground motion. hazard will be required.
7. Response Spectra. The response spectra
For Low Consequence Category dams and for
used for the hazard assessment should be
preliminary studies of higher Consequence
designed to allow for the response of the

Category dams it may be appropriate to conduct

dam and appurtenant structures at the dam.
an initial assessment based on existing PSHA
Conditional mean spectra may be used to
from nearby dams. Depending on the dam, its
represent the design response spectra for the
characteristics and whether its important
dam site.
potential failure modes are seismic loads

8. Minimum magnitude earthquake. The

related; e.g. liquefaction, a decision can then be
hazard assessment should be carried out for
made as to whether a site specific PSHA is
earthquake magnitudes Mw 5 and more.
However, under some circumstances,
smaller magnitude earthquakes may form In some situations for High and Extreme
the lower limit. With masonry dams, slab Consequence Category dams a staged approach
and buttress dams, older concrete dams and may be appropriate with a less detailed PSHA
structural concrete components of dams, as described above for Significant Consequence
Mw 4 earthquake should form the lower Category dams used initially, and the more
limit. detailed assessment carried out if it is
9. Quantification and reporting epistemic recognised that the seismic hazard is critical.
uncertainty. The epistemic uncertainty in Similarly for Significant Consequence Category
the true value of the hazard due to epistemic dams it may be appropriate to start with an
uncertainty in the earthquake source (1) and initial assessment based on an existing PSHA
ground motion models (4) should be from nearby dams, and then conduct to more
quantified using the fractiles of the hazard. detailed site specific assessment if it is
The report should include median (50th recognised that the seismic hazard is critical.

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 10
2.0 Assessment Of Design Seismic Ground Motions And Analysis Method

It should be noted that PSHAs become dated as Table 2.1 the OBE is often accepted as a
new methods are developed and data bases loading which has a 10% chance of being
improved. It is unlikely that a PSHA more than exceeded in a 50 year period, or an annual
about 5 years old will be reliable. The advice of probability of exceedance of 1 in 475. However
the Seismologist who carried out that study higher or lower annual probability loadings may
should be sought to advise if a new study is be adopted depending on the Owners appetite
required. for risk, and the criticality of the dam and
appurtenant structures or tailings dam.
2.6 Selection of Design Seismic Ground
Motion - Deterministic Analysis
Approach Safety Evaluation Earthquake SEE
Operating Basis Earthquake OBE
The recommended Safety Evaluation
The selection of the OBE is a matter for the Earthquake ground motions are shown in Table
Owner to consider in consultation with the 2.1.
Consultant and other stakeholders. As shown in
Table 2.1: Recommended deterministic analysis seismic design ground motions.
Dam Consequence Operating Basis Earthquake Safety Evaluation Earthquake

(1 )
Category OBE SEE ( 2 )
Extreme Commonly 1 in 475 AEP up The greater of:
Consequence to 1 in 1,000 AEP Ground motion from MCE on known
Category Dams active faults ( 3 )
Probabilistic ground motion
Extreme : 1 in 10,000 AEP ( 4 )

High A, B and C Commonly 1 in 475 AEP up Probabilistic ground motion ( 5 ) ( 6 ) :

Consequence to 1 in 1,000 AEP High A: 1 in 10,000 AEP

Category Dams High B: 1 in 5,000 AEP

High C: 1 in 2,000 AEP
Significant Commonly 1 in 475 AEP Probabilistic ground motion ( 5 ) :
Consequence 1 in 1,000 AEP

Category Dams

Low Consequence Commonly 1 in 475 AEP Probabilistic ground motion ( 5 ) :

Category Dams 1 in 1,000 AEP
(1) Owner and other Stakeholders to determine in consultation with the Consultant.
(2) A factor of safety should be applied to estimated stresses and deformations within the dam and its
foundations to give a low likelihood of the dam failing given the SEE loading.
(3) Active faults are as defined in Section 2.2. See C2.6 for discussion on active faults.
(4) 85th fractile.
(5) Median, 50th fractile.
(6) The adoption of these SEE for High B and High C Consequence Category dams may prove in
some particular cases not to provide an acceptable level of risk in accordance with ANCOLD Risk
Management Guidelines. It is therefore recommended that some level of risk assessment be
undertaken in these cases before adopting the AEP’ stated in the table. If it cannot be demonstrated
that an acceptable level of risk would be achieved then an AEP of 1 in 10,000 should be adopted.

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 11
2.0 Assessment Of Design Seismic Ground Motions And Analysis Method

2.7 Selection of Design Seismic For time-history analyses, use at least 3

Ground Motion - Risk Based ground motion records. Generally,
Analysis Approach accelerograms (acceleration vs time) will
As discussed in Section 2.3, for risk based be used although some analysis methods
assessments the seismic ground motions prefer velocity/displacement vs time. Use a
should cover the complete range of specialist Seismologist to develop
feasible loads up to and above the synthetic accelerograms (or
deterministic SEE values. velocity/displacement vs time where
required) if recorded accelerograms
2.8 Modelling Vertical Ground appropriate to the AEP earthquake are not
Motions available. Use amplitude scaling or
The vertical component of ground motions spectral matching to produce
should be included in the time-history accelerograms that will produce a response
accelerograms and used in dynamic spectrum that would approximately match
analyses for embankment and concrete the design response spectrum over the
dams. range of frequencies equivalent to the
range of natural frequencies of interest for
For pseudo-static analyses of concrete the dam.

dams the vertical ground motions should
be estimated using the method described in 2.10 Earthquake Aftershocks
the Commentary. Consider whether the dam, or appurtenant
structures, which are critical to operation
Vertical ground motions are often not of the dam post-earthquake are likely to be
incorporated in simplified deformation damaged sufficiently by the main
analyses for embankment dams. If required earthquake ground motion so as to be more
the methods described in the Commentary vulnerable to further damage by
should be used. earthquake aftershocks. If so, seek advice
2.9 Selection of Response Spectra and from the Seismologist on the likely
magnitude, focal location, and seismic

Time-History Accelerograms
ground motions at the dam site, and
A response spectrum will be required for analyse the effects on the dam and
each of the earthquake ground motion appurtenant structures. It also needs to be
cases (annual exceedance probabilities considered how the safety of the dam will

(AEP)) to be used in analysing the dam. be managed in the period through the
The response spectra should be prepared aftershocks until repairs can be completed.
by specialist Seismologists. The response
spectra are to be site specific, and for the 2.11 Seismic Ground Motions from
damping factor applicable to the dam. For Earthquakes Induced by the
example, response spectra for concrete Reservoir
dams with linear elastic response will The ground motion from Reservoir
usually be for a 5% damping factor Triggered Earthquakes (RTE) should be
(viscous damping). This linear elastic considered in the seismic hazard
analysis ignores the hysteresis damping assessment for new dams, either risk
with inelastic behaviour unless a non- based, or deterministically. This would be
linear inelastic analysis is carried out, a specific request to the Seismologist for
which is the most accurate to capture the advice on whether RTE should be
true behaviour of the dam or structures, considered and if so what the loadings
refer to the Commentary for additional would be. Refer to the Commentary for
details regarding damping. details.

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March
2017 12
3.0 Assessment Of Embankment Dams For Seismic Ground Motions

3 ASSESSMENT OF EMBANKMENT · The seismicity of the area in which the dam

DAMS FOR SEISMIC GROUND is sited and the assessed design earthquake.
MOTIONS · Foundation materials and topographic
conditions at the dam site.
3.1 Effect of Earthquakes on Embankment · The type and detailed construction, and
Dams natural period of the dam.
Earthquakes impose ground motions which · The water level in the dam at the time of the
result in additional loads on embankment dams earthquake.
over those experienced under static conditions.
The earthquake ground motion is of short The amount of site investigation, design, and
duration, cyclic and involves motion in the additional construction measures (over those
horizontal and vertical directions. Earthquakes needed for static conditions) will depend on
can affect embankment dams by causing any of these factors, the consequences of failure, and
the following: whether the dam is existing or new.

(a) Settlement and longitudinal and transverse There are four main issues to consider:
cracking of the embankment, particularly
near the crest of the dam. · Deformations induced by the earthquake
(settlement, cracking) and the effects on dam

(b) Liquefaction or loss of shear strength due to
increase in pore pressures induced by the freeboard.
earthquake in the embankment and its · The potential for liquefaction or strain
foundations. softening of saturated or nearly saturated
(c) Instability of the upstream and downstream sandy and silty soils and gravels with a sand
slopes of the dam if the seismic loading and silt matrix in the foundation, and
leads to strength loss e.g. from liquefaction, possibly in the embankment, and how this
within the embankment or foundation affects deformations during the earthquake
sufficient to result in post-earthquake factors and stability immediately after the
of safety less than 1.0. earthquake.
(d) Reduction of freeboard due to settlement or · The zoning and design of the dam,

instability which may, in the worst case, particularly the provision of filters, to
result in overtopping of the dam. prevent or control internal erosion of the
(e) Differential movement between the dam and the foundation, and provision of
embankment, abutments and spillway zones with good drainage capacity (e.g. free

structures leading to transverse cracks. draining rockfill).

(f) Transverse cracking in which internal · For tailings dams using upstream or
erosion and piping may develop. centreline construction, the potential for
(g) Differential movements on active faults liquefaction of loose to medium dense partly
passing through the dam foundation. saturated tailings where perched water tables
(h) Overtopping of the dam by seiches induced or nearly saturated zones may exist above
in the reservoir in the event of large tectonic the measured phreatic surface.
movement in the reservoir basin.
(i) Overtopping of the dam by waves due to 3.2 Defensive Design Principles for
earthquake induced landslides into the Embankment Dams
reservoir from the valley sides. The general philosophy is to apply logical,
(j) Damage to outlet works passing through the common-sense measures to the design of the
embankment leading to leakage and dam, to take account of the cracking, settlement
potential piping erosion of the embankment. and displacements which may occur as the result
of an earthquake. These measures are at least as
The potential for such problems depend on: important (probably more so) as attempting to
calculate accurately the deformations during

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 13
3.0 Assessment Of Embankment Dams For Seismic Ground Motions

earthquake. The most important measures which saturated, so they will be dilatant and not
can be taken are: liquefy.
(e) Avoid founding the dam on the strain
(a) Provide ample freeboard, above normal weakening clay soils and completely
operating levels, to allow for settlement or weathered rock, or rock with the potential to
slumping or fault movements which displace strain weaken Post- earthquake stability can
the crest. For example, one might adopt a be an issue if the earthquake causes even
narrow spillway with large flood rise and relatively minor movements which can take
large freeboard instead of a wide spillway the foundation strength from peak to
with small flood rise and thus usually a residual effective stress strength. Clay soils
lower freeboard, provided the costs were with high clay size fraction, and mudrocks
similar. are potentially strain weakening.
(b) Use well designed and constructed filters
downstream of the earthfill core (and There are a number of other less important
correctly graded rockfill zones downstream measures:
of the face for concrete, asphalt and other
membrane faced rockfill dams) to control (f) Use a well-graded filter zone upstream of
erosion if the core (or face) is cracked in the the core to act as a crack stopper, possibly
only in the upper part of the dam. The

earthquake. Filters should be taken up to the
dam crest level, so they will be effective in concept is that, in the event that major
the event of large crest settlements, which cracking of the core occurs in an earthquake,
are likely to be associated with transverse this filter material will wash into the cracks,
cracking. and prevent the core from eroding further by
(c) Provide ample drainage zones to allow for the crack stopper filtering against the
discharge of flow through possible cracks in downstream rockfill or filter, thereby
the core. For example provide that at least limiting flow and preventing enlargement of
part of the downstream zone is free draining the crack. If well designed filters are
or that extra discharge capacity is provided provided downstream, the upstream filter is
in the vertical and horizontal drains for an of secondary importance.

earthfill dam with such drains. In this regard (g) Flare the embankment core at abutment
some embankment dam types are inherently contacts, where cracking can be expected, in
more earthquake resistant than others. In order to provide longer seepage paths. Just
general the following would be in order of as (or more) important is to consider the
detailing of the contact with concrete walls

decreasing resistance:
o Concrete face rockfill. and the provision of filters downstream of
o Central core earth and rockfill. the contacts.
o Sloping upstream core earth and (h) Provide special details if there is likelihood
rockfill. of movement along faults or shears in the
o Earthfill with chimney and foundation.
horizontal drains. (i) Site the dam on a rock foundation rather
o Zoned earth-earth rockfill. than soil foundation (particularly if it is
o Homogeneous earthfill. potentially liquefiable or subject to strain
o Upstream construction tailings and weakening), where the option is available.
other hydraulic fill. (j) Use well graded (densely compacted)
(d) Avoid, densify, drain (to be non-saturated) sand/gravel/fines or highly plastic clay for
or remove potentially liquefiable materials the core, rather than clay of low plasticity (if
in the foundation or in the embankment. the option is available) (Sherard, 1967). The
Filters, rockfill and other granular materials former is more readily filtered by the
in the embankment should be well downstream zones, and the latter more
compacted if they are likely to become resistant to erosion than clays of lower

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 14
3.0 Assessment Of Embankment Dams For Seismic Ground Motions

When assessing an existing dam, the use of less than the available freeboard before
these “defensive design” measures is seldom the earthquake for the seismic load being
practical (except in remedial works). However, considered and the deformation are
it is useful to gauge the degree of security the much less than the width of filter,
existing dam presents by comparing it with this allowing for the uncertainty of the
list. Where the dam fails to meet many or most method, accept that the likelihood or
of these features, particularly (a) to (e), this may failure by overtopping and loss of filter
be a better guide to the fact that the dam may function is negligible for that seismic
not be very secure against earthquake than a lot ground motion.
of analysis. 2.2.If the deformations estimated by the
screening or database methods are
Dams which have well designed and constructed greater than this, use one or more of the
filters, have adequate stability against normal simplified methods to estimate
loads, and do not have liquefiable or strain deformations. If the estimated
weakening zones or foundations, will be likely deformations are significantly less than
to withstand the loading from even very large the available freeboard and filter width,
earthquakes with only minor deformations. allowing for uncertainty in the method
for the seismic load being considered,
3.3 Seismic Deformation Analysis of

accept that the likelihood of failure by
Embankment Dams
overtopping and loss of filter function is
Embankment dam engineers have recognised for negligible for that seismic ground
many years that when considering the effects of
seismic ground motions, it is deformations, not 2.3.If the deformations estimated by the
stability which should be assessed. Most simplified methods potentially threaten
embankments will under large seismic ground the dam with excessive settlement,
motions, yield during part of the loading cycle, opening of transverse cracks that present
resulting in permanent deformations. However a significant piping risk or shearing
that does not mean the dam has “failed” across filters for the maximum seismic
provided the deformations are tolerable. ground motion being considered, either

There are a number of methods available for use risk based methods to assess whether
estimating the deformations which may occur in remedial works or for a new dam, design
embankments and their foundations during and changes to reduce deformations are
post seismic loading. These vary considerably in required; or use an appropriate advanced

their degree of sophistication and time to do the numerical method to estimate

analyses. In view of this a staged approach to deformations and then use risk based
estimating deformations is recommended as methods to assess whether remedial
follows: works are required.
Repeat this as required for the full range
1. Assess whether the embankment or its of seismic loading being considered.
foundation are susceptible to liquefaction 3. For embankments and foundations subject to
using the methods described in Section 3.4. liquefaction and / or to significant pore
2. For embankments and foundations not pressure build up or strain weakening:
susceptible to liquefaction or experiencing 3.1 Assess post-earthquake factors of safety
significant pore pressure build up or strain using the limit equilibrium method. If
weakening: factors of safety for all reasonable lower
2.1.Use one or more of the screening or bound estimates of the liquefied strength
database methods to estimate the and allowing for pore pressure build up
deformation. If the estimates of and strain weakening in other zones are
deformations are much less than what is greater than 1.1 accept that the
tolerable; e.g. crest settlements are much likelihood of failure will be small and

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 15
3.0 Assessment Of Embankment Dams For Seismic Ground Motions

probably tolerable subject to the quality for these the deformations are at best
and extent of information available. This approximate, and controlled by the quality of
means that for mean estimates of the data input into the analyses, and the
liquefied strength the post-earthquake limitations of the methods of analysis.
factor of safety is likely to be > 1.3 or
1.5. The more sophisticated methods require expert
3.2 If factors of safety are lower than input to the analysis and selection of properties
described above, estimate deformations and should not be carried out other than by
using the simplified deformation analysis experienced persons.
method allowing for the range of
3.4 Assessment of the Effects of
liquefied strengths and other properties.
Liquefaction in Embankment Dams
Use these to assess the likelihood of
and Their Foundations
failure taking account of the very
approximate nature of these estimates. 3.4.1 Overall Approach
3.3 For High and Extreme Consequence When assessing the likelihood of failure of an
Category dams, regardless of the results embankment dam by loss of freeboard due to
of (3.1); and for other dams where liquefaction of the foundation soil, consider:
factors of safety are lower than described

in (3.1), carry out static numerical (i) The likelihood that the soils are
analyses of the post-earthquake susceptible to liquefaction or cyclic
condition for a range of liquefied softening.
(ii) The probability, given the earthquake
strengths and other properties. Use these
to assess the likelihood of failure taking seismic ground motion, that liquefaction
account of the approximate nature of occurs.
these estimates. (iii)Given liquefaction occurs, whether the
crest settles sufficiently to lose
For cases where the estimated likelihood of freeboard, taking account of the
failure are intolerable, either design remedial reservoir level at the time of the
works or for a new dam modify the design so earthquake.

the residual likelihoods of failure are tolerable, (iv)Given freeboard is lost, whether breach
or use advanced numerical methods to make occurs.
more refined estimates of deformations. Allow (v) Do this for both the upstream and
for the range of liquefied strengths and other downstream slopes of the embankment

properties. Use these to assess the likelihood of and for the varying geotechnical
failure taking account of the still approximate conditions which may apply over the
nature of these estimates. length of the embankment and its
Also consider interim actions if necessary to
(vi)If deterministic methods are being used
reduce the risk until remedial works are
for the assessment, it would be common
undertaken. to assume that the reservoir is at full
The extent to which an embankment is analysed supply level.
for deformations should be consistent with the
This assessment should allow for the modelling
Consequence Category and size of the dam. In
of uncertainties in estimating the seismic hazard,
many cases it may be better to design and
the likelihood of liquefaction, the liquefied
construct remedial measures or for new dams
strength, and deformations which may result.
modify the design than continuing to do more
These are by their nature very approximate. The
and more sophisticated analyses.
extent to which this is modelled will depend on
There may however be situations where the the Consequence Category of the dam, and
more sophisticated methods are warranted. Even whether the outcomes of the assessment are

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 16
3.0 Assessment Of Embankment Dams For Seismic Ground Motions

clear cut or marginal. These are more readily 3.4.3 Methods for Identifying Soils which
modelled if a risk based approach is being are Susceptible to Liquefaction
followed. (a) Geology and age of the deposit.
The following Sections provide guidance on Soils which are most susceptible to liquefaction
matters relating to liquefaction. The wording is are non-plastic or very low plasticity fills
somewhat prescriptive for the sake of brevity, including mine tailings and dredged fills, and
but they are not a standard and well qualified alluvial, fluvial, marine and deltaic soils.
and experienced practitioners may choose to Residual soils are generally likely to have a low
adopt alternative methods if that is appropriate susceptibility,
for the conditions they are assessing.
There is some indication that Pleistocene and
3.4.2 Definitions and the Mechanics of older soils may be more resistant to liquefaction
Liquefaction than Holocene soils but the evidence for this is
Liquefaction – All phenomena giving rise to a not sufficient to be relied upon for dam
reduction in shearing resistance and stiffness, engineering. This is discussed further in Section
and development of large strains as a result of 3.4.4 and in the Commentary.
increase in pore pressure under cyclic or
monotonic (static) loading of contractive soils. The age of the deposit should be allowed for

tailings dams and dredged fills as discussed in
Initial liquefaction – is the condition when Section 3.4.4.
effective stress is momentarily zero during
(b) Soil gradation, plasticity, moisture content
cyclic loading.

Flow liquefaction – is the condition where there Use well established methods such as those
is a strain weakening response in undrained described in the Commentary to assess whether
loading and the in-situ shear stresses are greater a soil is potentially liquefiable. The methods
than the steady state undrained shear strength. should be used with all required data inputs
including in-situ moisture content, Atterberg
Temporary liquefaction – is the condition limits, and fines content.

where there is a limited strain weakening

response in undrained loading; at larger strain Where Cone Penetration Test (CPT) data is
the behaviour is strain hardening. available use also CPT based methods such as
described in the Commentary. CPT or CPTU
Cyclic liquefaction – is a form of temporary based methods should not be relied upon alone.

liquefaction, where the cyclic loading causes

shear stress reversal and an initial liquefaction Use at least two and preferably more of the
(zero effective stress) condition develops well-established methods to assess the
temporarily. likelihood the soils are potentially liquefiable. A
suggested approach for doing this is given in the
Cyclic mobility – is a form of temporary Commentary.
liquefaction where the shear stresses are always
greater than zero. 3.4.4 Methods for Assessing Whether
Liquefaction may occur
Cyclic softening – as a term used to describe the Use well established methods such as those
reduction of shear strength and stiffness of clays described in Section C3.4 of the Commentary
and plastic silts under cyclic loading. subject to the qualifications detailed in the
It is recommended that all those involved in
liquefaction assessments read the documents If a new consensus method or updates of the
referenced in the Commentary to familiarise methods described herein are published in
themselves with the mechanics of liquefaction. refereed journals, these methods should be

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 17
3.0 Assessment Of Embankment Dams For Seismic Ground Motions

Take account of the fact that for some of the use strengths between the best estimate and
referred methods the deterministic methods for lower bound strengths.
estimating Cycle Resistance Ratio (CRR) are for
15% probability of liquefaction, that there are Use methods published in refereed journals such
significant uncertainties in the factors used as those referenced in Section C3.4.5 of the
within these methods and as a result, soils which Commentary.
plot within the margins of liquefiable and non-
3.4.6 Methods for Assessing the Post-
liquefiable soils should be assigned some
Earthquake Strength of Non-
likelihood of being liquefiable.
Liquefied Soils in the Embankment
These uncertainties are greater for soils below and Foundation
about 15 metres from the surface. Allow for the effects of pore pressure build up
during cyclic loading in saturated non liquefied
Take account also of the uncertainty in the non plastic strata, and for cyclic softening of
geotechnical model, and variability of the saturated sensitive plastic strata.
Standard Penetration Test (SPT) and CPT data,
and the uncertainty of the earthquake loading. Allow for cracking and potential weakening of
compacted fills.
Where the Ks and Ka values are critical to the

assessment of liquefaction, and / or the Use methods such as those described in the
liquefiable strata are below 15 metres it may be Commentary.
necessary to seek expert advice and use more 3.4.7 Site Investigations Requirements and
advanced methods to assess liquefaction. Development of Geotechnical Model
Do not use a mix of factors from the different of the Foundation
methods because the authors rely on their own The key requirements are:
approach when developing the data bases upon (a) Review of available data relating to the
which the methods are developed. embankment foundation.
The effects of ageing of the soil deposit may be (b) Develop a preliminary geotechnical

taken account of as detailed in the Commentary. model and plan any additional site
This should be done for tailings and dredged investigations required.
fills particularly if they are recently deposited. (c) Carry out site investigations using the
methods described in the Commentary.
3.4.5 Assessing the Strength of Liquefied (d) Refine the geotechnical model using all

Soils in the Embankment and the available data

Use both “critical state” and “normalised Details are given in the Commentary.
strength ratio” methods and apply equal 3.4.8 Liquefaction Analysis
weighting to each method. Follow the procedure detailed in the
Do not allow for any increase in the liquefied Commentary.
residual strength beneath a berm which is to be 3.4.9 Assessment of the Likelihood and the
added to improve post-earthquake stability. If a Effects of Cracking of Embankment
berm has already been constructed carry out Dams Induced by Seismic Ground
SPT and CPT in the liquefiable soils below the Motions
berm and use these data to assess the liquefied When assessing the likelihood of failure for
strength. cracking under earthquake loading leading to
Take account of the fact that there are internal erosion and piping consider:
significant uncertainties in these methods, and

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 18
3.0 Assessment Of Embankment Dams For Seismic Ground Motions

(i) The probability, given the earthquake, only part of the length of the embankment may
settlement occurs resulting in transverse require remedial works; e.g. if only part is
cracking. founded upon liquefiable soil. It is not
(ii) Given it occurs, whether it will persist to uncommon to use a combination of methods;
below the reservoir level at the time of e.g. to carry out treatment on the downstream
the earthquake or before repairs can be but do nothing upstream and tolerate the risks
carried out. posed by upstream deformations. The ground
(iii)Given it does, whether erosion will improvement method used will depend on the
initiate along the crack, whether filters nature and depth of the liquefiable soils, the
will prevent erosion continuing, whether techniques available, and cost.
erosion progression and a breach forms.
3.5.2 Embankment Dams not Subject to
This is seldom a dominant failure mode because Liquefaction
“normal” and “flood” load conditions tend to The following are the most common remedial
dominate internal erosion and piping failure measures which may be required:
modes. It should be noted that dams which are
susceptible to cross valley differential settlement (a) Provision of filters or upgrade of
may not require large earthquakes to produce existing filters in the upper part of the
dam. This is usually done to reduce risks

transverse cracking.
of internal erosion and piping to
3.5 Methods for Upgrading Embankment tolerable levels for flood and normal
Dams for Seismic Ground Motions loading, and the upgrade for earthquake
loads is achieved at the same time.
3.5.1 General Approach (b) For the very few dams where freeboard
The method or methods suitable for upgrading a is insufficient, the embankment may
dam and its foundation for seismic ground have to be raised, usually along with
motions will be site specific, and where raising for flood.
applicable will require a thorough understanding (c) Where there are strain weakening soils
of the liquefaction mechanics, extent, and the such as high clay content high plasticity
consequences for the dam. It will also depend on

over-consolidated clays in which strains

the objectives, probably measured in residual may localise under earthquake
risk terms. deformations, it may be necessary to add
a stabilizing berm.
The remedial measures may consist of one or

more of the following: 3.5.3 Embankment Dams Subject to

(a) Do nothing and accept the potential
The following are the most common remedial
damage and risks.
measures which may be required:
(b) Adding filters and possibly raising the
crest level of the embankment so the (a) Construction of a stabilising berm, most
consequences of deformations and the commonly founded upon the potentially
resulting risks are reduced to tolerable liquefiable soil after it has been treated
levels. by ground improvement. Alternatively
(c) Modifying the liquefiable soil in the remove the potentially liquefiable soil
foundation (and the embankment if from the foundation of the berm.
applicable), and / or constructing (b) Provision of filters or upgrade of
stabilizing berms to limit deformations existing filters in the upper part of the
to tolerable levels. dam to reduce the risks of internal
erosion and piping to tolerable levels.
These works may be required only on the
(c) Where freeboard is insufficient, the
downstream of the embankment, or on both
embankment may have to be raised.
upstream and downstream. Quite commonly

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 19
3.0 Assessment Of Embankment Dams For Seismic Ground Motions

Section C3.5.3 in the Commentary includes a

summary of ground improvement methods, their
limitations and some design details.
3.6 Flood Retarding Basins
The assessment of retarding basins under
seismic ground motions should be considered
similar to the methods described above for
embankment dams. It is to be noted that it is
likely that a retarding basin will be empty when
subjected to seismic loading conditions and
therefore the resulting risks associated with the
earthquake are likely to be low. However it is
important that the condition and function of the
retarding basin post-earthquake be considered
with regard to the potential filling of the
retarding basin following an earthquake.


ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 20
4.0 Assessment Of Concrete Dams For Seismic Ground Motions

4 ASSESSMENT OF CONCRETE shaking, some of these structural elements may

DAMS FOR SEISMIC GROUND progressively fail. In many circumstances
MOTIONS however, this will not necessarily lead to
failure of the dam – simply to a redistribution
4.1 Effect of Earthquakes on Concrete of loads/stresses to other parts of the dam. It is
Dams important to appreciate this when analysing the
It is important to recognise the different dam. Failure may be either structural failure of
behaviours that gravity, arch and buttress dams the various elements of the dam or global
exhibit during and after earthquake shaking. sliding.

With gravity dams, cracking can occur which is 4.2 Defensive Design Principles for
likely to initiate on the upstream face or at the Concrete Dams
upstream heel, and propagate along the base of The use of defensive design principles for new
the dam or along lift surfaces. This cracking concrete dams as well as for the upgrading of
would then produce increased existing concrete dams is advised.
deformations/movement in the dam, allow
increased uplift pressures to develop post- Defensive design principles to ensure that these
earthquake which could be exacerbated by criteria are met, should address the dam

partial or fully blocked drainage due to the structure, the interface between the dam and its
deformations and could damage/rupture foundations, and the dam’s foundations. In
passive or active ground anchors due to general terms, the dam or the dam’s remedial
excessive deformations, thereby reducing post- works should be designed / assessed such that:
earthquake stability. Gravity dams are most
(a) Sliding during seismic excitation is
likely to fail post-earthquake by sliding along
the foundation interface, through the ideally prevented, although some
deformation may be inevitable under
foundations or along a weak lift surface.
Failure can also be caused by over-stressing at large seismic ground motions. Sliding
resistance should be sufficient,
a sudden change in geometry in the upper part
of the dam that leads to cracking severe enough potentially with reduced shear

resistance along the sliding surface, to

to cause toppling/sliding of the upper part of
the dam. meet the post-earthquake stability
Arch dams generally have more complex (b) The dam is stable against overturning
post-earthquake at all levels of

modes of vibration than gravity dams and some

over-stressing is more likely to occur above the earthquake ground motion up to SEE,
base of the dam (as well as along the base of allowing for cracking and increased
the dam). There is likely to be some uplift pressure in the foundations
redistribution of stresses within the dam to the caused by deformations during the
dam’s abutments. Failure is likely to occur due earthquake. It is to be noted that under
to sliding along the foundation interface, post-earthquake conditions that the
through the foundation or along any weak concrete dam may sit in this condition
perimetric joint. Failure could also be due to for some time, this should be taken into
greater horizontal thrusts and shears being consideration in the assessment of the
transferred to gravity abutment blocks leading outcomes of the analysis.
to instability in those abutment blocks. Failure (c) Any cracking of the dam or in the
could also occur of the abutment rock mass foundations will not lead to
which supports the arch (hoop) forces from the uncontrolled leakage.
dam. (d) Hydraulic outlet structures are not
damaged to the extent that they either
Buttress dams may comprise a number of allow uncontrolled loss of water which
structural elements (especially slab and buttress might lead to collapse of the dam or
dams). During the course of earthquake

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 21
4.0 Assessment Of Concrete Dams For Seismic Ground Motions

they cannot be used to lower the storage strains and consequently much lesser
if necessary. stresses due to earthquake loading.
· Detail internal features such as galleries
To achieve the above, the following defensive so they do not give rise to stress
design principles are recommended: concentrations which will could lead to
excessive cracking.
· Proportion the cross section of a dam
with no sudden changes in shape or 4.3 General Principles of Analysis for
section stiffness. Seismic Ground Motions
· Keep any superstructures on the dam See Commentary for some general design
crest to a minimum. principles and refer ANCOLD (2013).
· Avoid sudden changes in the abutment
profile which would give rise to stress 4.4 Seismic Structural Analysis of
concentrations, or design to Concrete Dams
accommodate the differential
displacements which are likely to occur 4.4.1 Material Properties
at these locations.
· Account for geological features in the Concrete
Refer to the ANCOLD (2013) for information

foundations which would control the
dam's sliding stability. If there are such on selecting appropriate strength and stiffness
features, then suitable means of parameters for concrete, when analysing
stabilising the dam to account for them concrete dams. Consideration should be given
should be employed. This is especially though to any increase in strength or stiffness
true of arch dams, since the interaction due to the short term and high frequency nature
of dam thrusts and foundation of earthquake loading. This is discussed in the
characteristics is determinative of the Commentary.
whole dam-foundation stability.
In assessing existing concrete dams, it will be
· Provide sufficient internal and
necessary to not only investigate the strength
foundation drainage, and see that drain
and stiffness of the concrete but also the

holes are of large enough diameter that

strength of lift surfaces.
they will remain operative particularly
if some sliding is expected under Foundations
seismic ground motions. It is vital that The design of concrete dams requires a

the drains keep working or that the dam thorough knowledge of the foundation
is stable post-earthquake with the drains conditions.
not operative.
· Undertake adequate investigations of The design process should include a properly
the dam’s foundations and of the lift funded geotechnical investigation of the rock
joints to be able to determine the foundations.
strength parameters of these features.
· If post-tensioned ground anchors are The investigations should provide a detailed
required for stabilising a dam then geological model of the dam foundations as
ensure that they are un-bonded (except described in ANCOLD (2013). The model
for their anchorage length) - this will should be used as the basis for estimating the
allow strains due to any overloading design strength, compressibility, design uplift
resulting from earthquake loading to be pressures for the foundations and kinematically
taken over the entire free length of the feasible failure mechanisms.
anchor rather than over a very short The design team should include specialists with
length either side of a crack in the dam considerable experience in dam foundation
or foundations – strains taken over the investigation and design, engineering geology
entire free length results in much lesser and rock mechanics. The investigation team

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 22
4.0 Assessment Of Concrete Dams For Seismic Ground Motions

should develop the geological model and work Seismic

with the dam engineering design team to
estimate foundation design parameters. General
In preliminary, simplified studies where only a
The investigations should be done in stages horizontal response spectrum is supplied, the
with the development of the dam design vertical response spectrum may be generated
feeding back into subsequent stages of the by scaling the horizontal spectral acceleration
geotechnical investigations. using the multipliers given in the Commentary.
For more detailed analyses us the methods to
A critical issue is whether there are continuous, apply the ground motions described in Section
or near continuous, unfavourably oriented C2.9.
discontinuities in the foundation, such as
bedding surfaces, bedding surface shears, joints Inertia
including stress relief (sheet) joints, faults and Consider the mode shape of the dam and the
shears. The geological investigations should associated natural frequency, to determine the
assess the foundation for such features and, if acceleration of the dam at various levels within
present, the design should make allowance for the height of the dam. Ensure that sufficient
them. modes of vibration contribute to the

determination of inertia loads.
The shear strength of the foundation should be
determined as outlined in Section 5.4 of Hydrodynamic
ANCOLD (2013). The compressibility should
be estimated as detailed in the Commentary. Gravity and Buttress Dams:
Pressure distributions which assume the dam is
4.4.2 Loads rigid and the water incompressible may be used
for preliminary studies, especially for low Static height dams. Other methods as described in the
Static loads on concrete dams shall be in Commentary should be applied to more
accordance with ANCOLD (2013). complex dam geometry (e.g. fully or partly
sloping upstream face) and for high dams.
R Uplift
The pre-earthquake uplift pressure distribution Arch Dams:
estimated in accordance with ANCOLD (2013)
should be assumed to apply during the Review the fundamental frequency of the dam

earthquake. with an empty storage compared to the

fundamental resonant frequency of the storage.
The post-earthquake uplift pressure distribution Where the ratio of the two natural frequencies
should be determined considering the cracking is near unity, it is necessary to consider the
developed during earthquake shaking. flexibility of the dam and the compressibility of
Consideration may be given to the the water in the storage. In this case is will be
effectiveness of internal/foundation drains necessary to consider the storage in any finite
following the earthquake. If cracking element model of the dam (e.g. model the
propagates past the line of drains, the capacity storage (all or part) using fluid elements) in
of the drains should be assessed to check order to determine hydrodynamic pressures on
adequacy for handling potentially larger the upstream face of the dam. See the
discharges due to water travelling along the Commentary for further information on
crack. calculating hydrodynamic pressures for arch
dams. Silt
Consider any potential liquefaction of silt when Dynamic Earth Pressures
estimating silt loads in a post-earthquake Dynamic earth pressures from silt, fill or
analysis. abutments should be calculated based on
geotechnical principles as discussed in the
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 23
4.0 Assessment Of Concrete Dams For Seismic Ground Motions

Commentary for Section 6.6.4. Pseudo-static or accelerations through overburden is not likely
pseudo-dynamic procedures will be sufficient to be an issue as it may be for embankment
in most cases. dams founded on soil) and for the relevant
damping factor and design earthquake peak Load combinations ground acceleration.
Refer to ANCOLD (2013) for applicable load
combinations. For analyses in the time domain, obtain a
minimum of 3 sets (preferably 5 sets) of time-
Consideration should be given to the behaviour histories (two orthogonal horizontal time-
of a concrete dam during an earthquake when histories and one vertical time-history in each
the dam is empty or near empty. This is likely set).
to be important in gravity dams which have
been strengthened using post tensioned anchors For analyses in the time domain either direct
(which are likely to be located near the dam’s integration or modal superposition methods
upstream face). Cracking propagating from the may be used to determine stresses in the dam.
downstream toe of the dam could prejudice the The former is the more accurate but requires
stability of the dam during subsequent flood considerably greater computer resources
loading of the dam. (computer memory and run time). The latter
will be more conservative as it uses a statistical

As noted in ANCOLD (2013), consideration approach (e.g. square root, sum of the squares)
also needs to be given to the possible loss of to combine the maximum stresses from the
stabilising force from damaged post-tensioned various modes. Consequently, it ignores the
cables by shearing due to sliding displacement sign of the maximum stress for a particular
or tension overload. mode.
For arch dams especially, temperature effects Undertake structural seismic analysis in a
can have a significant effect on the stress hierarchical manner. That is, analysis should
distribution within the dam. However, in most start using simplified linear elastic methods and
arch dams built with vertical contraction joints, then progress as required, to more sophisticated
the contraction joints tended to be grouted methods as given in Table 4-1. The

during winter. Therefore winter represents the Commentary gives more details on
temperature stress neutral condition. In applicability of the methods.
summer when the dam deflects upstream due to
concrete expansion, the earthquake stresses The extent to which a concrete dam is analysed

induced in the dam are likely to be less than in should be consistent with the Consequence
winter. However, each dam should be Category and size of the dam. In some cases it
considered on its merits and at least a may be better to design and construct remedial
qualitative assessment made of how the dam is measures than continuing to do more and more
likely to perform during earthquake loading in sophisticated analyses (e.g. non-linear
combination with different temperature analysis). There may however be situations
conditions. where the more sophisticated methods are
warranted. The more sophisticated analysis
4.4.3 Methods Available and When to Use methods should not be carried out other than by
Them experienced persons.
Analysis will be either in the frequency domain
where an earthquake response spectrum is
used, or in the time domain where time-
histories of velocity, acceleration or
displacement are used.

For analyses in the frequency domain, obtain a

response spectrum for the site (in the case of
concrete dams, attenuation or amplification of
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 24
4.0 Assessment Of Concrete Dams For Seismic Ground Motions

Table 4-1 Hierarchy of Seismic Analysis Methods for Concrete Dams

i. Simplified, linear elastic methods using a site specific response spectrum on a 2D model of
the dam for pseudo-static cantilever type stability analysis – generally only applicable to a
concrete gravity dam.
ii. Simplified, linear elastic methods using a site specific response spectrum on a 2D finite
element model (FEM) of the dam and foundations – applicable to a concrete gravity dam or
for the upstream/downstream direction of a concrete buttress dam.
iii. Linear elastic methods using a site specific response spectrum on a 2D or 3D finite element
model (FEM) of the dam and foundations – applicable to a concrete gravity dam (2D
satisfactory for flat sloped abutments or buttress dams considering only
upstream/downstream direction); 3D required for steep abutments), and for arch or buttress
dams (considering full 3D action).
iv. Simplified, linear elastic methods using site specific time-history records of ground
acceleration or velocity on a 2D or 3D finite element model (FEM) of the dam and
foundations – applicable to a concrete gravity dam (2D satisfactory for flat sloped abutments
or buttress dams considering only upstream/downstream direction); 3D required for steep
abutments, and for arch or buttress dams (considering full 3D action).
v. Non-linear methods using site specific time-history records of ground acceleration or velocity

on a 2D finite element model (FEM) of the dam and foundations so that cracking can be
simulated – applicable to a concrete gravity dam or for the upstream/downstream direction of
a concrete buttress dam.
vi. Non-linear methods using site specific time-history records of ground acceleration or velocity
on a 3D finite element model (FEM) of the dam and foundations so that cracking can be
simulated – applicable to a concrete gravity, an arch dam or for a concrete buttress dam.
The level of complexity of the analysis may be linked to the dam’s Consequence Category as indicated
in Table 4-2.

Table 4-2 Analysis Method Appropriate for Consequence Category


Analysis Methods Consequence Category

Low Significant High and Extreme
2D linear elastic, simplified response spectrum, Yes Yes Yes (1)
cantilever analysis - gravity dam only ((i) in Table

2D linear elastic FEA, site specific response Yes Yes
spectrum-gravity dam & buttress dam in u/s-d/s
direction ((ii) in Table 4-1)
2D or 3D linear elastic FEA, site specific response Yes
spectrum- all concrete dams ((iii) in Table 4-1)
2D or 3D linear elastic FEA, site specific Yes
accelerograms- all concrete dams ((iv) in Table 4-1)
2D non-linear FEA, site specific accelerograms - Yes
gravity dam & buttress dam in u/s-d/s direction ((v)
in Table 4-1)
3D non-linear FEA, site specific accelerograms – all Yes
concrete dams ((vi) in Table 4-1)
(1) Consideration needs to be given by the dam Owner in consultation with the Consultant as to whether
a 2D analysis is considered sufficient for a High and Extreme Consequence Category dam. In most
cases a higher level of analysis would be required.

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 25
4.0 Assessment Of Concrete Dams For Seismic Ground Motions

4.4.4 Analysis in the Frequency Domain unacceptable DCR, carry out a non-
See Commentary for details. linear time-history analysis to estimate
the extent of sliding/rocking that the
4.4.5 Analysis in the Time Domain dam undergoes. Assess this against the
See Commentary for details. estimated maximum sliding/rocking that
the dam can undergo. This assessment
4.5 Approach to Analyses and Acceptance should consider excessive permanent
Criteria leakage that may occur through the dam
or its foundations, and the effect on
4.5.1 During the Earthquake
internal drains and post tensioned Gravity Dams: anchors.
The following are the steps in the analyses and Arch Dams:
acceptance criteria. The extent of analyses will
Carry out a 3D linear elastic finite element
depend on the Consequence Category of the
analysis (FEA) of the dam and its foundations
dam, and the purpose of the analysis; e.g.
using response spectra, in the first instance. Go
concept versus detailed design. For a detailed
to more sophisticated 3D linear elastic FEA
design or assessment of a High or Extreme
using time-histories and 3D non-linear FEA
Consequence Category dam linear elastic time-

using time-histories as might be dictated by the
history or non-linear -history time analyses
results of the less sophisticated analyses and the
should be carried out. It is likely that a response
dam’s (and foundation’s) geometry/properties.
spectrum analysis will give overly conservative
See Commentary for details.
results. This will be important if the response
spectrum analysis gives an extent of cracking Buttress Dams:
during earthquake shaking that would lead to The requirements are that for both
inadequate post-earthquake stability. upstream/downstream and cross valley ground
1. Carry out pseudo-static analysis using motions:
5% damping factor. If factors of safety · The structure has satisfactory factors of
are ≥ minimum required by ANCOLD safety against sliding and overturning as

(2013) then behaviour during required for concrete gravity dams or that
earthquake loading is satisfactory. permanent deformations are tolerable.
2. If factors of safety are < required by · The face slabs are not dislodged and satisfy
ANCOLD (2013) carry out linear elastic normal structural reinforced concrete

response spectrum analysis or linear requirements;

elastic time-history analysis. If stresses · The buttresses do not fail in buckling when
using 5% damping factor indicates assessed using normal reinforced concrete
stresses ≤ dynamic strengths, then design principles;
behaviour during earthquake loading is · Sufficient struts remain intact such that the
satisfactory. strength of the buttresses is not
3. If (2) indicates stresses > dynamic compromised.
strengths, then carry out analysis with
response spectrum modified for 10% 4.5.2 Post-Earthquake
damping factor. If stresses > dynamic Gravity Dams:
strengths then undertake linear elastic
Regardless of the method of analysis used to
time-history analysis to estimate
analyse the dam during earthquake shaking,
Demand Capacity Ratio (DCR) and
carry out a post-earthquake stability analysis.
cumulative time for non-linear
This analysis should consider damage to the
behaviour. Assess according to relevant
dam occurring during the ground motion (e.g.
USACE manual (e.g. USACE EM
disruption of drains, failure of post tensioned
anchors) and changes in uplift pressure
4. If (3) indicates that the linear elastic
distribution due to cracking caused during
analyses undertaken indicate
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 26
4.0 Assessment Of Concrete Dams For Seismic Ground Motions

earthquake ground motions. Reduced or

residual shear strength parameters due to strain
weakening should be used where appropriate
within the foundations and at lift joints. The
minimum post-earthquake factor of safety for
sliding should satisfy ANCOLD (2013) Table
6.1 and 6.2 criteria. Arch Dams

Consider extent of cracking at base of dam
caused by earthquake shaking in determining
redistribution of stresses in the dam and in
determining post-earthquake sliding resistance.
Consider extent of cracking in both the arch and
any gravity abutment sections in considering
the change of loading on the abutment sections.
Use residual shear strength parameters within

the dam and the foundation where appropriate
according to the requirements of ANCOLD
AF Buttress Dams
Carry out post-earthquake analysis following
the principles for gravity dams as discussed in
The minimum factor of safety for sliding and
overturning should satisfy the requirements of

ANCOLD (2013) for gravity dams.

4.5.3 General
See Commentary.

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 27
5.0 Assessment Of Tailings Dams For Seismic Ground Motions

5 ASSESSMENT OF TAILINGS DAMS Tailings dams and their potential impacts must
FOR SEISMIC GROUND MOTIONS be effectively managed throughout their life,
from design through operation to closure and
5.1 Some General Principles post closure. The post closure period may be
This Section of the Guideline supplements the very long; e.g. ANCOLD (2012a) suggest 1000
ANCOLD Guideline on Tailings Dams years.
(ANCOLD, 2012a). Those involved with
tailings dams should refer to that Guideline. The flow diagram below illustrates a cyclical
return step during operation indicating possible
Where there are conflicting data in the two modifications to ensure that the dam achieves its
Guidelines this Guideline represents ultimate goal - sustainable closure with minimal
ANCOLD’s position. post closure risks and associated costs.

Where there are additional State regulations for These considerations can impact on the seismic
tailings dams these regulations must be design requirements of tailings dams as
complied with. described below.

O p e ra te M a in ta in & D e c o m m is s io n
D e s ig n C o n s tr u c t A fte r c a re
& M o n ito r M o d ify & C lo s e
Tailings dams may be designed and constructed There is also a greater susceptibility to
by a number of methods (ANCOLD, 2012a): liquefaction and strain weakening leading to
large deformations and potentially slope
· Downstream, usually incorporating a instability which needs to be taken into account.
starter dam.

· Centreline, where part of the 5.2 The Effects of Earthquakes on Tailings

embankment is built over the tailings Dams
beach. For downstream construction these are as for
· Upstream, where most of the conventional embankment dams as detailed in

embankment is built over the tailings Section 3.1. If the water storage is kept remote
beach. from the dam by good water management
practices the implications of cracking of the dam
A combination of these methods may be used; potentially leading to internal erosion and piping
e.g. downstream construction for the majority of may be less than for a conventional dam.
the dam, with small raises using upstream or
centreline construction. For tailings dams constructed by centreline and
upstream construction methods there will be a
If downstream construction methods are used greater emphasis on liquefaction and strain
there are few if any differences between such weakening, and potential for large deformations,
dams and conventional water retaining dams internal erosion and piping in cracks resulting
from the seismic design viewpoint. from deformations and slope instability.
However, if centreline or upstream construction 5.3 Defensive Design Principles for Tailings
is used there are a number of the methods Dams
described in this Guideline for conventional
These are as for conventional embankment
dams which are not applicable. These are
dams as detailed in Section 3.2. For tailings
discussed below.
dams constructed by centreline and upstream

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 28
5.0 Assessment Of Tailings Dams For Seismic Ground Motions

construction methods there will be a greater (d) Where feasible use “integrated waste
emphasis on the potential for liquefaction and management” principals to construct
strain weakening and slope instability. conservative embankments from mine waste
Particular care and conservatism is required if
upstream construction methods are proposed for (e) Provide internal drainage systems to
High and Extreme Consequence Category lower phreatic surface and reduce the level of
tailings dams if they are constructed of tailings saturation of tailings;
materials which are potentially liquefiable. This
is due to the difficulty of maintaining good (f) Maximise the density and reduce
construction practices throughout the life of the saturation by well-planned and managed tailings
dam, and the uncertainty of predicting the discharge aimed at minimising thickness of
liquefied strength of the tailings and post- fresh tailings discharge layers and maximising
earthquake deformations. If upstream evaporation and drying shrinkage;
construction methods are used it should only be
(g) Maximise the density and reduce
by involving Consultants expert in these saturation by mechanical working of the tailings
matters, using conservative assumptions and
using purpose built equipment such as
requiring confirmation of assumed parameters Amphirolls (Archimedes Screw Tractors):– note

during construction.
that this method has been able to successfully
Upstream construction may reasonably be used compact tailings to non-liquefiable conditions
for small raises of tailings dams constructed by equivalent to “dry-stacking”;
downstream methods where it can be
(h) Use dry-stacking methodology including
demonstrated with a high degree of confidence
filter-press technology to dewater tailings and
that the potential effects of liquefaction and / or
potentially allow placement and compaction of
strain weakening of the tailings can be
tailings to non-liquefiable density or saturation
accommodated with a large margin of safety.
Where upstream construction is used, this must
5.4 Design Seismic Ground Motions and
recognise that such dams are not suitable for

Analysis Method
water storage other than the minimum amount
required for achieving decant water quality. Design seismic ground motions for tailings
Water storage limitations should be set, and dams during operation are as for conventional
stringently adhered to during operation and for dams and are given in Table 3.2. The

the life of the dam, taking into account the lack Consequence Category should be determined as
of filters for this type of dam and the potential detailed in ANCOLD (2012), Table 1. It should
deformation, which should ensure no contact of be noted that for tailings dams the impact on the
stored water to the perimeter embankment. natural environment may be a controlling factor.

Where centreline undertaken, specific defensive The design seismic loads for tailings dams post
design principles include: closure should be the SEE. The SEE should be
re-evaluated at the time of closure to allow for
(a) Use conservative design methods any development in understanding of seismic
including estimates of the extent of liquefaction hazard and for any changes in the consequences
and residual strength parameters for stability of failure, and that allowance should be made
analysis; for potential increases in the PAR and hence
consequences post closure.
(b) Include buttresses constructed of high
strength, non-liquefiable materials as part of the
outer shell design;
(c) Avoid water storage on the dam surface;

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 29
5.0 Assessment Of Tailings Dams For Seismic Ground Motions

5.5 Seismic Deformation Analysis of

Tailings Dams not Subject to
These are as for conventional embankment
dams as detailed in Section 3.3.

For tailings dams constructed by centreline and

upstream construction methods there will be a
greater emphasis on liquefaction and strain
weakening. None of the “screening”, “database”
or “simplified” methods are applicable to
tailings dams where liquefaction or other
significant strain weakening may occur.
Deformations may be estimated by the methods
described for static and advanced numerical
deformation analyses as described in Sections
C3.3.4.3 and C3.3.4.4 of the Commentary.

5.6 Assessment of the Effects of
Liquefaction in Tailings Dams and
Their Foundations
These are as for conventional embankment
dams as detailed in Section 3.4.
Allow for the effects of age of the tailings
deposit as described in Section 3.4.3 and Section
C3.4.3 in the Commentary.

5.7 Methods for Upgrading Tailings Dams


for Seismic Loads

These are as for conventional embankment
dams as detailed in Section 3.5.

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 30
6.0 Assessment Of Appurtenant Structures For Seismic Ground Motions

6 ASSESSMENT OF APPURTENANT requirements of the component and should be

STRUCTURES FOR SEISMIC determined by the Owner and Consultant. It
GROUND MOTIONS may be required for appurtenant structures that
are critical to the operation and safety of the
6.1 The Effects of Earthquakes on dam following an earthquake, e.g. low level
Appurtenant Structures outlet works that have the ability to draw down
A number of subsidiary structures associated the reservoir, that these structures maintain
with a dam are essential for the dams operation. function post-earthquake loading conditions
For these structures it is important that their with a high degree of confidence. To achieve
functional and structural integrity is retained in this may require them to be designed for the
the event of a notable earthquake. This is SEE.
particularly the case where the appurtenant
structure may be required to release water from When designing and assessing appurtenant
the reservoir in a controlled manner to lower the structures it is important to note that these are
storage following an earthquake. It is therefore hydraulic structures and the assessment needs to
necessary that not only the intake/outlet be conducted accordingly. Australian Standards
structures and their gates and valves remain such as AS1170 (Structural Design Actions) are
serviceable but also that there is proper access typically not appropriate for these structures.

to these structures. Bridges and roads may need When assessing appurtenant structures
to remain in a sound state after an earthquake consideration also needs to be given to other
depending upon their importance. external factors that could influence the
The most important factors in considering the operation of the structure such as rock fall
earthquake resistant design of an appurtenant impacts, restricted access for operation, loss of
structure are: power, etc.

· Whether or not failure of such a structure This following sections provide the information
could lead to loss of control of the reservoir for the various types of appurtenant structures
following an earthquake, including:

· Where the structure is required for post- · Defensive design principles;

earthquake operation to lower or maintain · Performance criteria for the Operating Basis
the reservoir level so that the structure Earthquake (OBE) and the Safety
maintains its functionality or allow repairs. Evaluation Earthquake (SEE); and

Generally, appurtenant structures should be · Recommended analysis procedures.

such that: 6.2 Intake Towers
· They maintain their normal operating 6.2.1 General
condition after an OBE
· They are not damaged to an extent where Dam intake/outlet facilities typically comprise
they could allow sudden or uncontrolled of the intake/outlet structure, intake/outlet
loss of water from the storage for ground tunnel, exit structure, access bridge (at times)
motions up to the SEE. and operating equipment including pipework,
· Following ground motions up to the SEE, valves, generators, electrical control panels, etc.
appurtenant structures should operable to Many intake/outlet towers, termed intake
an extent that the dam is able to pass floods towers from herein, are either free standing on
while repairs are carried out. an enlarged base or foundation mat placed on
The level of assessment and design ground the reservoir bottom, or are deeply founded
motions of critical appurtenant structures is through bedrock or soil, away from the dam.
subject to the risk assessment and operational Others are embedded within earthfill dams, or

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 31
6.0 Assessment Of Appurtenant Structures For Seismic Ground Motions

structurally tied to the upstream face of intake tower due to deflections occurring
concrete dams. Towers built within during the seismic event.
embankment dams may interact dynamically
with the surrounding materials. Reservoir water 6.2.3 Performance Requirements
surrounds most towers, sometimes up to a OBE: Static and dynamic loads to induce
significant height. Hence they are subjected to maximum concrete and steel
hydrodynamic interaction effects. Some contain reinforcement stresses within the elastic
inside water, which also affects seismic region (i.e. limited amount of
response. reinforcement yielding) for strength
design purposes and the tower and its
6.2.2 Defensive Design Principles base is to remain stable.
The following defensive design principles are
suggested for new structures and upgrades to SEE: Significant amount of reinforcement can
existing structures: yield and damage to the tower may
· Site the intake tower on a foundation where occur. The tower and its base and should
there are no geological features that could not be damaged to an extent that it leads
lead to uneven displacement or to an uncontrolled release of water from
deformations during the earthquake event. the reservoir. Intake towers required for

· Where possible, the intake tower should be the emergency release of water
socketed into rock to assist with preventing following an earthquake event should
the tower from sliding. remain functional.
· Where possible, install grouted dowels in 6.2.4 Analysis Procedures
the rock foundation as a redundancy against
uplift and rocking. The analysis methods for intake towers are
· Avoid so far as is practicable sudden described here in general terms only. It is
changes in profile that could give rise to recommended that the reader refer to ICOLD
stress concentrations. (2002), USACE (2003b), USACE (2007) and
the Commentary for additional detailed
· Provide horizontal reinforcement designed
information on the analysis procedures. The

to prevent vertical reinforcement from

buckling and to confine concrete when it is more sophisticated methods require expert input
in compression. The horizontal to the analysis and should not be carried out
reinforcement should be placed on the other than by experienced persons.

outside of the vertical reinforcement. The general issues and potential modes of
· In the design of towers, practices should be failure that need to be examined in the seismic
adopted that ensure ductile behaviour while response of intake towers include the following:
suppressing brittle failure modes including:
meet minimum reinforcing steel · Flexural displacement demands exceeding
requirements such that the nominal strength the flexural displacement capacity;
moment is equal to 1.2 times that cracking · Shear demands exceeding shear (diagonal
moment; provide adequate confinement at tension) capacity;
splice locations and plastic hinge regions: · Shear demands exceeding the sliding shear
provide anti-bucking hoops/ties in plastic capacity; and
regions; provide adequate splice and anchor · Moment demands exceeding the
lengths; avoid locating splices in inelastic overturning capacity (rocking).
regions; and provide direct and continuous
loads paths. There are a number of methods available for the
· For access bridges to intake towers provide assessment of intake towers, these are provided
appropriate support mechanisms designed to below ranging in increasing level of
prevent damage to both the bridge and the complexity. It is to be noted that care is to be
taken when selecting models with increasing

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 32
6.0 Assessment Of Appurtenant Structures For Seismic Ground Motions

levels of complexity and is to be based on the time consuming. In this instance it needs to be
judgement of experienced personnel. This is considered whether remedial measures would
particularly the case with non-linear analysis be more beneficial than further analysis.
where the assessment can often be complex and

Table 6-1: Hierarchy of Seismic Analysis Methods for Intake Towers

Strength Assessment
i. Response Spectrum: The response spectrum analysis is adequate for towers whose responses
to earthquakes are within the linear elastic range. If it is determined that the tower remains
within its elastic range under the OBE and meets the ductility requirements for the SEE, then
the tower is considered acceptable.
ii. Linear Time-History Analysis: Applicable if tower exceeds requirements of the Response
Spectrum Analysis. If the demand capacity ratios, cumulative duration of bending moment
excursions and extent of reinforcement yielding meet those described in the Commentary then
the tower is considered acceptable.
iii. Non Linear Analysis: The nonlinear analysis of towers can be complicated and time
consuming. In this instance it needs to be considered whether remedial measures would be

more beneficial than further analysis.
Stability Assessment - Sliding
i. Response Spectrum: The OBE is considered an unusual condition and a factor of safety of 1.3
is required. The SEE is considered an extreme condition and a factor of safety of 1.1 is
ii. Linear Time-History Analysis: The tower should meet the requirements described in Section
6.2.3 for the OBE and SEE. Within this analysis, the stability is maintained and sliding does
not occur if the factor of safety is greater than 1.
iii. Non Linear Analysis: In the nonlinear time-history analysis an assessment can be made of the
total permanent sliding displacement of the tower. It then needs to be assessed whether the
permanent displacements meet the criteria for the SEE described in Section 6.2.3.

Stability Assessment – Rotational

i. Tipping Potential Evaluation: The tower may start to tip and start rocking during an
earthquake. The assessment of the tipping potential for the tower, either a rigid or flexible
tower, is provided in USACE (2007). If it is determined that no tipping occurs, ie. the structure

does not break contact with the ground, then the tower is considered stable. If tipping occurs
then the rocking block analysis should be conducted.
ii. Rocking Block Analysis: USACE (2007) and the Commentary provide the procedures for
assessment of the tower rocking. The tower should meet the requirements described in Section
6.2.3 for the OBE and SEE.
Note. For sliding in the foundation the principles of selection of the strength of the foundation
and factors of safety should be as detailed in ANCOLD (2013).

6.3 Spillways (conveyance structures such as floor slab with

walls connecting the inlet structures to the
6.3.1 General terminal structure), and the terminal structure
Typically, spillways are constructed of mass or (hydraulic-jump, stilling basin, flip bucket,
reinforced concrete. Seismic loads often control impact structure, etc.). Each of these spillway
the design of such structures. Spillway components is covered in the following section.
component structures can be grouped into three
general classes, including inlet structures (inlet
and/or crest structures including gates), chutes

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 33
6.0 Assessment Of Appurtenant Structures For Seismic Ground Motions

6.3.2 Defensive Design Principles spillways should be operable to an extent

that the dam is able to pass floods while
For defensive design principles for mass and
repairs are carried out.
structurally reinforced components of spillways,
refer to Section 4.2. 6.3.4 Analysis Procedures
6.3.3 Performance Requirements General
OBE: Spillways should maintain their normal Any large mass spillway structures that retain
operating condition after an OBE. the reservoir should be designed in accordance
with Section 4 – Concrete Dams. For all other
SEE: Spillways, including gates that retain
spillway structures, Table 6-2, taken from
permanent storage at the time of an SEE
ICOLD (2002) with some additions, provides
should not fail to an extent where water
some guidance on the recommended analysis
is released in an uncontrolled manner.
methods for the components of spillway
Following ground motions up to the SEE

Table 6-2: Analysis method of spillway component structures

Spillway Components Usual Approaches Recommended models
Inlet and crest Morning glory drop Response spectrum 3D
structures inlet structures Linear time-history
Overflow structures: Pseudo-static 2D plane strain of plain
Straight ogee crests, Using elastic stress or
Labyrinth, Fuse gates foundation 3D
Siphon structures Response spectrum 2D or 3D
Fuse plug structures: Deformation analysis 2D
zoned embankment Newmark method or

(refer Section 3)
Chutes Conveyance Pseudo-static 2D plane-strain or plane-

structures: Using elastic stress

Floor slab and foundation
connecting walls
Terminal Structures Hydraulic jump Pseudo-static 2D
Stilling basin Using elastic
Flip bucket
Impact structures

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 34
6.0 Assessment Of Appurtenant Structures For Seismic Ground Motions

Additional inertia loading from operational 6.4.4 Analysis Procedures

equipment, piers, etc. needs to be considered If spillway gates are located on the top of a
in the assessment of spillway structures. For concrete dam, the spillway gates will need to
spillway piers, the effect of the bridges and be analysed as an integral part of the dam.
operating equipment needs to be considered Amplification of the ground accelerations
in the assessment of the piers. could be significant (in some cases there may
6.4 Spillway Gates be a two – to fivefold increase in magnitude).

6.4.1 General Seismic loads transmitted from spillway gates

to trunnion pins and trunnion blocks should
Spillways often incorporate gate systems that also be accounted for when designing or
retain permanent storage and / or fulfil dam assessing a gated crest structure. These loads
safety functions. These gates can, at times, be can be significant due to their concentrated
used to allow the controlled release of water in nature.
a potential dam safety emergency, including
post earthquake. Hydrodynamic loads are typically assessed as
described in Section, and applied as
Spillway gate types typically include crest added masses to a model. When using this

mounted radial gates, orifice radial gates, approach, the following items need to be
vertical lift wheel gates and flap gates. considered as they can impact on the
hydrodynamic loads applied to the gates:
6.4.2 Defensive Design Principles
· The depth of the water against the gate;
The following defensive design measures are
· The depth of water against the full spillway
· Provide sufficient flexibility and details in
· The fundamental frequency of the gate
the gate system to accommodate expected
compared to the fundamental resonant
differential movements during an
frequency of the storage; and
earthquake event.
· The position of the gate relative to the
· Evaluate the reliability of the operating

upstream face of the spillway.

system, including appropriate access, for
operation post earthquake, refer to Sections With the development of numerical modelling,
6.8 and 6.9 for additional details. the use of fluid or acoustic elements within a
· The gate drive systems should not fail or

three dimensional model is now possible and

distort; such as drive shafts and hydraulic can be adopted, particularly for structures with
cylinders. a complex geometry. For further details on
seismic induced loads on spillway gates refer
6.4.3 Performance Requirements to USBR (2011).
OBE: Spillways gates and their operating
gear should maintain their normal The complete gate system should be
operating capability after an OBE. considered, from incoming power supply,
electrical components, backup supplies, hoist
SEE: Spillways gates that retain permanent design and gate design, emergency bulkheads
storage at the time of an SEE should and cranes, operator access.
not fail to an extent where water is
released in an uncontrolled manner. The effect of dynamic amplification on critical
The SEE should not cause gates to components (eg hydraulic cylinders) may need
distort or gate piers to permanently to be considered.
displace to an extent where the gates
become jammed. The hoist system or a
backup arrangement should remain

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 35
6.0 Assessment Of Appurtenant Structures For Seismic Ground Motions

6.5 Outlet Works - Water Conduits, Gates · Consider rock falls that could occur as a
and Valves result of an earthquake and potentially
damage the pipeline or penstock.
6.5.1 General · For pressurised systems consideration
Water conduits such as pipelines, penstocks, needs to be given to the hydrodynamic
tunnels and low-level outlets may be required forces that can be generated during an
for reliable, controlled, rapid emptying of the earthquake, refer to the Commentary.
reservoir. Water conduit design should be such · Ensure the gate actuation systems can
that it does not lead to failure or compromise withstand vibration. Hydraulic cylinders
the functioning of the dam and its foundation. are vulnerable to vibration. Rope supported
In addition, in the case of water supply gates may be susceptible to vibration under
reservoirs in populated areas, the safety and certain conditions and position.
operability of the outlet pipelines, gates and
valves become significant factors affecting the 6.5.3 Performance Requirements
maintenance of drinking water supplies as well Outlet Conduits
as water to fight-fires and assisting with post-
earthquake recovery functions. OBE: Static and dynamic loads to induce
concrete and steel reinforcement

6.5.2 Defensive Design Principles stresses which satisfy AS3600 Concrete
The following defensive design measures are Structures.
SEE: Conduit not to collapse or rupture.
· Provide sufficient flexibility in the outlet Collapse could lead to an undermining
conduit system to accommodate expected and subsequent failure of the
differential movements during an embankment. Rupture could cause
earthquake event. piping or destabilise an embankment by
· Weak zones, faults and active fault a marked increase in pore pressure.
crossings should be avoided where
possible for outlet conduits. If such areas Pipelines and Penstocks

cannot be avoided, then design details

should be utilised which can accommodate OBE: Static and dynamic loads to induce
displacement and differential movements. steel stresses which satisfy AS4100
· Tunnel plugs can be incorporated as part of Steel Structures.

the conduit design to prevent uncontrolled

SEE: Pipelines and penstocks required for
releases of water should damage or conduit
emergency releases or post-earthquake
failure occur.
recovery not to collapse or rupture.
· Where possible, pipelines or penstocks,
including supports and anchor blocks, Gates and Valves
should be founded on suitable rock, soil, or
stabilised soil which is capable of OBE: All gates and valves to maintain their
minimising differential movement and normal operating capabilities.
settlement due to seismic ground motion.
SEE: Emergency closure and regulating gates
· Pipelines or penstocks should not be
and valves (especially low level release
founded on low-density, non-plastic soils
valves) to maintain operating capability
that are subject to liquefaction or high
– the storage may need to be quickly
levels of strain that could cause damage
lowered if parts of the dam are
even during a moderate earthquake.
damaged and need remedial works or
· Consider the differential loadings that can
relief of hydrostatic loads.
occur on conduits which pass through
different zones in an embankment.

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 36
6.0 Assessment Of Appurtenant Structures For Seismic Ground Motions

6.5.4 Analysis Procedures · Provide sufficient drainage provided

The complete outlet system should be behind the retaining wall. It is vital that the
considered, including power supply, electrical drains keep working or that the retaining
components, backup supplies, valve and gate wall is stable post-earthquake with the
design and operation, emergency bulkheads drains not operative;
and cranes, operator access. · If post-tensioned ground anchors are
required for stabilising a retaining wall
Outlet Conduits design so that they are un-bonded (except
for their anchorage length), refer to Section
Use methods detailed in the Commentary. A 4 for additional details on anchors;
2D analysis is usually sufficient for the seismic · For new structures and upgrades design
assessment of an outlet conduit. and specify the backfill to the retaining
wall so that the backfill materials are not
Pipelines and Penstocks
liquefiable as this can lead to significant
Consideration needs to be given to the deformations of the backfill material,
potential of rocking of pipelines and penstocks increasing the potential for opening a gap
within outlet conduits and the potential between the wall and backfill resulting in
amplification effects that can occur. piping and increasing the loads on the

retaining wall.
6.6 Retaining Walls · For retaining walls that retain
embankments, the wall be sloped on the
6.6.1 General
embankment side to maintain positive
Retaining walls are often critical components contact with the embankment following
to a spillway or dam structure. They include deformations that may occur due to the
gravity, semi gravity and non-gravity walls earthquake.
with seismic loads often controlling the design
of such structures. During an earthquake, a 6.6.3 Performance Requirements
retaining wall can be subjected to dynamic soil OBE: Retaining walls should maintain their
pressures caused by motions of the ground and

normal operating condition after an

the wall that need to be accounted for in the OBE. Static and dynamic loads to
design and assessment of the wall. For induce maximum concrete and steel
retaining walls that retain embankments, reinforcement stresses which are
consideration needs to be given to the within the elastic region.

performance of the backfill material as this has

the potential to lead to piping or large SEE: Walls which retain part of an
deformations, particular consideration needs to embankment dam shall not collapse or
be given to backfill materials that are deform to an extent that could lead to
potentially liquefiable. embankment failure due to piping or
breach due to significant deformations.
6.6.2 Defensive Design Principles
6.6.4 Analysis Procedures
The following defensive design measures are
recommended: Gravity retaining wall structures which if they
· Avoid sudden changes in the retaining wall fail could result in breach of the dam should be
profile which would give rise to stress designed in accordance with Section 4 –
concentrations; Concrete Dams. Cantilever retaining walls
· Site where there are no geological features which if they fail could result in breach of the
in the foundations which would decrease dam should be designed to the same principles
the retaining walls sliding stability - if but using design methods appropriate for such
there are, then suitable means of stabilising walls as described in the Commentary.
these features or accounting for them in the
design should be employed.
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 37
6.0 Assessment Of Appurtenant Structures For Seismic Ground Motions

The dynamic loads applied from the backfill 6.8 Mechanical and Electrical Equipment
material need to be included. Depending on the
magnitude of wall movements the backfill 6.8.1 General
material is said to be in yielding, non-yielding, For gates and outlet works to remain
or intermediate state. Further information on functional following an earthquake event any
the assessment of the backfill loads is provided mechanical and electrical equipment required
in USACE (2007) and in the Commentary. for the operation of the gates and valves also
need to remain operational. These items in
6.7 Parapet Walls existing dams are often not designed for
6.7.1 General seismic ground motions and can sometimes be
fragile and may be subjected to accelerations
Parapet walls are often used to increase the many times greater than the peak ground
height of embankment dams and are generally acceleration due to amplification of the ground
constructed using either cast-in-situ or precast motion. Depending on its resonant frequency,
concrete components. These walls are above equipment may be susceptible to dynamic
the full supply level of the dam and do not amplification resulting in high stresses. Some
retain permanent storage. ‘off the shelf’ equipment may not be suitable
for seismic loading and vibration.

6.7.2 Defensive Design Principles
· Provide sufficient flexibility in parapet 6.8.2 Defensive Design Principles
wall joint details to allow for deformations The following defensive design measures are
due to the earthquake event. recommended:
· Consider the potential for internal erosion · Anchor mechanical and electrical
and piping underneath the walls and if components to maintain their stability
required install suitable filter zones. during earthquake events and their
operability following the event. This is
6.7.3 Performance Requirements
typically relatively inexpensive to
OBE: Parapet walls should maintain their incorporate into new and existing

normal operating condition after an structures.

OBE. · The reliability of the power supply should
be evaluated. Consideration should be
SEE: Parapet walls can be expected to given to alternative power supplies for
deform and be damaged in a SEE. The critical components. Consider the

Owner and Consultant should consider possibility of landslides and structural

the potential for damage, how this may movement on cable runs and hydraulic
be minimised and details provided so pipework. Control buildings should be
that the dam will satisfy tolerable risk designed to withstand earthquake loading.
criteria resulting from floods following Consider fuel tanks, switchboards,
the earthquake and before the parapet transformers, cable trays and support of
wall can be repaired. starter batteries.
6.7.4 Analysis Procedures · Critical crane equipment should be
appropriately anchored to avoid
Refer to the Section 6.6: Retaining Walls for misalignment. Especially applicable to
analysis procedures. Consideration of the cranes supporting emergency bulkheads or
performance and deformation of the parapet required for emergency remedial works.
wall needs to be given as part of the · Hoist systems (shafts, hydraulic cylinders
earthquake assessment of the underlying dam and pipework etc.) should be appropriately
structure. Consider the amplification effects designed.
that can occur at the top of the dam.

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 38
6.0 Assessment Of Appurtenant Structures For Seismic Ground Motions

6.8.3 Performance Requirements support for the bridge from its

All mechanical and electrical equipment and bearings.
o Failure of the girders, supports,
components should meet the design
handrails, etc. that could interfere
requirements of the equipment that they
with the gate operating machinery or
operate. All equipment should be appropriately
interfere with opening of the gates.
anchored to prevent sliding, overturning or
o Failure of the beam seats that could
impact loading. All power facilities, including
interfere with the gate operation
electrical conduits and buildings that might
o Differential movement that could
house engine generator sets, should be
cause hydraulic lines or other power
designed to functionally survive the design
supply functions to fail.
seismic ground motion.
o Distortion that could interfere with
6.8.4 Analysis Procedures communication/control functions.
· For access bridges to intake towers the
A pseudo-static method of analysis is usually
appropriate support mechanisms should be
sufficient for the assessment of mechanical and
chosen to prevent or limit damage to both
electrical components. This needs to take into
the bridge and the intake tower due to
account that the accelerations may be greater
deflections occurring during the seismic

than the bedrock peak ground acceleration, and
event. This includes provision that
the potential for dynamic amplification of the
sufficient longitudinal and lateral
vibration in some equipment.
movement is available in the bridge
6.9 Access Roads and Bridges supports.
· For critical structures an assessment may
6.9.1 General need to be conducted on the potential for
In order to allow access to the dam for rock falls or landslides and the impact they
inspection or operation of outlet works may have on either the structure itself or on
following an earthquake event, the bridges and access to the structure. It may be
roads may need to remain in a sound state impractical to prevent rock falls, even in
events more frequent than the OBE, and

depending upon the importance of the dam and

its appurtenant structures. rock fall protection structures may be
There is a limit to what a Dam Owner has
control over in this instance as typically roads 6.9.3 Performance Requirements

and bridges to access dams are not owned or OBE: Access roads and bridges to the dam
maintained by the Dam Owner, this needs to and its appurtenant works that are
be considered in assessing the requirements for owned and maintained by the Dam
operation of appurtenant structures following Owner should so far as is practicable
an earthquake. remain passable after an OBE event.
Therefore, likely rock fall or landslip
6.9.2 Defensive Design Principles areas along the access roads may need
The following defensive design measures are to be checked for potential instability.
recommended: The impact of this on access roads and
· For bridges over spillways that are bridges that are not owned or
combined with or attached to the maintained by the dam owner
supporting structure for operating becoming impassable needs to be
mechanisms that open and close gates a considered in assessing the required
number of defensive design principles can operation of appurtenant structures
be incorporate to prevent the following following an earthquake.
from occurring:
o Distortion of the piers that could SEE: Access roads to the dam and its
cause binding of gates or removal of appurtenant works may become

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 39
6.0 Assessment Of Appurtenant Structures For Seismic Ground Motions

impassable after an SEE event. with AS5100 – Bridge Structures. For intake
However, for access roads that are tower access bridges, the analysis may be part
owned and maintained by the Dam of an overall dynamic analysis for an intake
Owner they should so far as is tower. The pier and bridge system should be
practicable be easily cleared or made examined for seismic ground motions in the
accessible if there are no alternative direction of the bridge access and
access routes available. Access bridges perpendicular to the bridge access. Loads
that are owned and maintained by the resulting from seismic ground motions may be
Dam Owner should remain capable of combined on a square root of the sum of the
carrying the design loads if there are no squares (SRSS) basis.
alternative access routes available. The
impact of this on access roads and 6.10 Reservoir Rim Instability
bridges that are not owned or If there are existing or potential landslides on
maintained by the Dam Owner the reservoir rim, assess the likelihood that the
becoming impassable needs to be slides will activate under seismic loads, and if
considered in assessing the required so, whether the slide debris will reach the
operation of appurtenant structures reservoir and at what velocity and volume.
following an earthquake. Then assess the potential for waves to be

generated by the landslide debris, and if so, the
For bridges over spillways these may effects on the dam and appurtenant structures.
be combined with or attached to the
supporting structure for operating
mechanisms that open and close gates.
For purposes of dam safety, such
bridges should not impede the ability to
open and close spillway gates
following the occurrence of the SEE.

6.9.4 Analysis Procedures (Bridges)


Bridges should be assessed using an

appropriate dynamic analysis of the pier and
bridge system and should be in accordance

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 40
Commentary - Table of Contents

C1 INTRODUCTION………...………………………………………… ........ ……………………44

C2.1 Earthquakes and Their Characteristics ............................................................................... 45
C2.1.1 Seismology for Earthquake Hazard Studies ................................................................. 45
C2.1.2. Earthquake Recurrence .............................................................................................. 51
C2.1.3. Seismic Ground Motion ............................................................................................. 55
C2.1.4. Earthquake Hazard in Australia .................................................................................. 60
C2.1.5. Reservoir Triggered Earthquake ................................................................................. 64
C2.2Terminology....................................................................................................................... 66
C2.3 General Principles for Selection of Design Ground Motion and Analysis Method.............. 66
C2.4 Description of a Probabilistic Seismic Hazard Assessment ................................................ 67
C2.5 Requirements of a Seismic Hazard Assessment ................................................................. 76

C2.6 Selection of Design Seismic Ground Motion - Deterministic Approach ............................. 90
C2.7 Selection of Design Seismic Loading- Risk Based Approach ............................................. 92
C2.8 Modelling vertical ground motions .................................................................................... 93
C2.9 Selection of Response Spectra and Time History Accelerograms ....................................... 93
C2.10 Earthquake Aftershocks ................................................................................................... 94
C2.11 Seismic Ground Motions from Earthquakes Induced by the Reservoir ............................. 95

C3.1 Effect of Earthquakes on Embankment Dams. ................................................................... 96

C3.2 Defensive Design Principles for Embankment Dams. ........................................................ 96
C3.3 Seismic Deformation Analysis of Embankment Dams. ...................................................... 96

C3.3.1 The Methods Available and When to use them............................................................ 96

C3.3.2 Screening and Empirical Database Methods ................................................................ 97
C3.3.3 Simplified Methods for Estimating Deformations during Earthquakes ........................ 99
C3.3.4 Post-earthquake Deformations for Liquefied Conditions ........................................... 100
C3.3.5 Advanced Numerical Methods for Estimating Deformations During and Post-
Earthquake for Non-Liquefied and Liquefied Conditions ..................................................... 101
C3.4 Assessment of the Effects of Liquefaction in Embankment Dams and Their Foundations 103
C3.4.1 Overall Approach...................................................................................................... 103
C3.4.2 Definitions and the Mechanics of Liquefaction ......................................................... 103
C3.4.3 Methods for Identifying Soils which are Susceptible to Liquefaction ........................ 104
C3.4.4 Methods for Assessing Whether Liquefaction may occur .......................................... 106

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 41
Commentary - Table of Contents

C3.4.5 Methods for Assessing the Strength of Liquefied Soils in the Embankment and
Foundation ........................................................................................................................... 112
C3.4.6 Methods for Assessing the Post-Earthquake Strength of Non-Liquefied Soils in the
Embankment and Foundation ............................................................................................... 114
C3.4.7 Site Investigations Requirements and Development of Geotechnical Model of the
Foundation ........................................................................................................................... 115
C3.4.8 Liquefaction Analysis ............................................................................................... 117
C3.4.9 Assessment of the Likelihood and the Effects of Cracking of Embankment Dams
Induced by Seismic Ground Motions .................................................................................... 117
C3.5 Methods for Upgrading Embankment Dams for Seismic Ground Motions ....................... 117
C3.5.1 General approach ...................................................................................................... 117
C3.5.2 Embankment Dams not Subject to Liquefaction ........................................................ 118
C3.5.3 Embankment Dams Subject to Liquefaction .............................................................. 118

C3.6 Flood Retarding Basins ................................................................................................... 119
C4.1 Effect of Earthquakes on Concrete Dams......................................................................... 120
C4.1.1 General ..................................................................................................................... 120
C4.1.2 ICOLD ..................................................................................................................... 121
C4.2 Defensive Design Principles for Concrete Dams ............................................................. 121
C4.3 General Principles of Analysis for Seismic Ground Motions ........................................... 121

C4.3.1 Gravity Dams............................................................................................................ 121

C4.3.2 Arch Dams ................................................................................................................ 121
C4.3.3 Buttress Dams ........................................................................................................... 122

C4.3.4 Other Types .............................................................................................................. 122

C4.4 Seismic Structural Analysis of Concrete Dams ................................................................ 122
C4.4.1 Material Properties .................................................................................................... 122
C4.4.2 Loads ........................................................................................................................ 123
C4.4.3 Methods Available and When to Use Them .............................................................. 124
C4.4.4 Analysis in the Frequency Domain ............................................................................ 126
C4.4.5 Analysis in the Time Domain .................................................................................... 128
C 4.5 Approach to Analyses and Acceptance Criteria .............................................................. 128
C4.5.1 During the Earthquake .............................................................................................. 128
C4.5.2 Post-Earthquake ........................................................................................................ 129
C4.5.3 General ..................................................................................................................... 130

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 42
Commentary - Table of Contents


C5.1 Some General Principles ................................................................................................. 131
C5.2 The Effects of Earthquakes on Tailings Dams ................................................................. 131
C5.3 Defensive Design Principles for Tailings Dams ............................................................... 131
C5.4 Design Seismic Loads and Analysis Method ................................................................... 131
C5.5 Seismic Deformation Analysis of Tailings Dams not Subject to Liquefaction .................. 132
C5.6 Assessment of the Effects of Liquefaction in Tailings Dams and Their Foundations ........ 132
C5.7 Methods for Upgrading Tailings Dams for Seismic Loads ............................................... 132
................................................................................................................................................ 133
C6.1 The Effects of Earthquakes on Appurtenant Structures .................................................... 133
C6.2 Intake/Outlet Towers ....................................................................................................... 133

C6.2.1 General ..................................................................................................................... 133
C6.2.2 Defensive Design Principles ..................................................................................... 133
C6.2.3 Performance Requirements ....................................................................................... 133
C6.2.4 Analysis Procedures .................................................................................................. 133
C6.3 Spillways ........................................................................................................................ 135
C6.4 Spillway Gates ................................................................................................................ 136
C6.5 Outlet Works - Water Conduits, Gates and Valves........................................................... 136

C6.5.1 General ..................................................................................................................... 136

C6.5.2 Defensive Design Principles ..................................................................................... 136
C6.5.3 Performance Requirements ....................................................................................... 136

C6.5.4 Analysis Procedures .................................................................................................. 136

C6.6 Retaining Walls ............................................................................................................... 136
C6.6.1 General ..................................................................................................................... 136
C6.6.2 Defensive Design Principles ..................................................................................... 136
C6.6.3 Performance Requirements ....................................................................................... 136
C6.6.4 Analysis Procedures .................................................................................................. 137
C6.7 Parapet Walls .................................................................................................................. 137
C6.8 Mechanical and Electrical Equipment .............................................................................. 137
C6.9 Access Roads and Bridges ............................................................................................... 137

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 43
Commentary – 1.0 Introduction

No commentary.


ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 44
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis
up to a few metres. Once ruptured, the fault is
C2 ASSESSMENT OF DESIGN SEISMIC a weakness, which is more likely to further
GROUND MOTIONS AND rupture in future earthquakes, so a large total
ANALYSIS METHOD displacement may build up from many
earthquakes over a long period of time. This
C2.1 Earthquakes and Their may eventually measure kilometres for thrust
Characteristics faults produced by compression, or hundreds
of kilometres for horizontal strike-slip faults
C2.1.1 Seismology for Earthquake Hazard such as the San Andreas fault in California.
As shown in Figure C2.1 the point on the
C2.1.1.1 Earthquake Mechanisms and fault surface where a displacement
Terminology commences is called the hypocentre or focus,
An earthquake is the sudden slip on a fault and the earthquake epicentre is the point on
that is produced when stress within the earth the ground surface vertically above the
builds up over a long period of time until it hypocentre.
eventually exceeds the strength of the rock,
which then fails and a break along a fault is

produced. It may take tens, hundreds or
thousands of years for the stress to build up
in a particular area, and it is then released in
seconds. Part of the energy is transmitted
away as seismic waves and part as heat.
The fault displacement in a particular
earthquake may vary from a few millimetres

Figure C2.1 Definition of earthquake terms.

The fault slip initially propagates away from propagating in both directions along strike
the hypocentre in a circular manner, but its and sometimes only in one direction.
eventual rupture extent is variable, sometimes Maximum energy release is not necessarily

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis
near the hypocentre, and is often not even The vertical principal stress at any depth
close to the hypocentre. usually has a value comparable with the
‘lithostatic pressure’, or ρgh, where ρ is the
The hypocentral distance from an earthquake density of the rocks above the point. The two
to a point is the three-dimensional slant near horizontal principal stresses can be
distance from the hypocentre to the point, higher or lower than this value, or one can be
while the epicentral distance is the horizontal higher and the other lower, depending on
distance from the epicentre to the point. which of σ1, σ2, or σ3 is vertical.
Tectonic stress within the earth is caused by If maximum principal stress σ1 is vertical, or
deformation that results from plate ‘normal’ to the Earth’s surface we get a
movement. The stress can be resolved into normal fault, resulting in a horizontal
three orthogonal principal stresses, the extension. If the minimum principal stress σ2
maximum, intermediate and minimum is vertical we get a reverse fault, resulting in
principal stresses, usually denoted by σ1, σ2, horizontal compression. If the intermediate
and σ3. principal stress σ3 is vertical a strike-slip fault
is produced, in which two blocks move
For crustal earthquakes, one of these
horizontally relative to each other.

principal stresses is usually near vertical, so
the other two will be near horizontal. The
stress directions determine the type and
orientation of the faulting.
Fault Plane

Extension Shortening

Normal Reversed Strike-Slip

Fault Fault Fault

Figure C2.2: Principal stress orientation and fault types (P: Maximum, B: Intermediate, T:

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis
A fault will only be active if its type and C2.1.1.2 The Fault Rupture Process
orientation is consistent with the principal The earthquake process is more complex
stress orientation, to within about 20 than the simple displacement between two
degrees. Ancient faults that are not oriented blocks shown in Figure C2.2.
appropriately are very unlikely to fail, but
those that are so oriented may be A fault is a weakness that has experienced
reactivated. Stress orientations may be many ruptures in the past, each one
inferred from earthquake mechanisms, or providing an incremental shift between the
obtained from the World Stress Map. blocks on either side. As tectonic
deformation takes place at a very slow but
In Australia, all measurements to date show fairly constant rate, the relative motion
the maximum principal stress is near between blocks is concentrated along the
horizontal, with a high level of stress fault, with bending that results in shear
compared with the lithostatic pressure. The strain, and the accumulation of shear strain
minimum principal stress is usually vertical energy in the blocks on both sides of the
so most faults are reverse. In some locations fault as shown in Figure C2.3.
the intermediate principal stress is near
vertical, so active faults at these places are Since the Earth’s surface is a free surface,

likely to be strike-slip. this will result in folding and a vertical
offset. This deformation happens very
Using traditional mining terms for normal slowly over a very long period of time. The
and reverse faults, the block that lies over folding has low amplitude relative to the
the dipping fault is called the hanging-wall horizontal extent, metres vertical over
block, and the one under the fault is the kilometres or tens of kilometres horizontal
foot-wall block. distance. Reverse faults give folding in the
hanging-wall block, similar to the folds
For strike-slip faulting, standing on one produced by subduction in the over-riding
block and facing the other, then observing plate, but without outer-rise folding on the
which way the other is moving (right or subducting plate. Small earthquakes occur

left), defines right- or left-lateral strike-slip throughout the folded region during
faulting. deformation.

Figure C2.3: Deformation associated with reverse faulting.

The main active reverse fault dips under the few earthquakes in the footwall block. Some
uplifted block, so epicentres of foreshocks foreshocks and aftershocks may also occur
and aftershocks that are on this fault extend above the main fault in the up-thrown block.
for tens of kilometres on the up-thrown
block side of the fault surface rupture, with

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis
Fault Rupture and generated seismic waves because faults are weaker than surrounding
unbroken rock, and are much more likely to
When the stress level exceeds the strength fail again when stress rebuilds. Earthquakes
of the fault at a point on the fault, it nucleate within the brittle zone of the
ruptures. The stored strain energy allows the shallow crust, which is typically about 10 to
rupture to propagate along the fault, despite 20 km thick. There is a brittle-to-ductile
considerable loss of energy due to friction transition at the base of this seismogenic
and some due to radiation of seismic waves. zone, typically at a depth of 10 to 20 km,
Propagation continues until there is caused by the increase in temperature with
insufficient strain energy to continue the depth in the earth. Because rocks near the
rupture, kinetic energy drops to zero, and surface are relatively weak in non-cratonic
the rupture displacement stops. regions of Australia (as defined in Clark et
al., 2011; 2012; Figure C2.20), there is
When the rupture occurs the release of shear
another brittle-to-ductile transition at
strain energy leads to the production of a
shallow depths, typically a few km, so that
shear wave (S wave) that travels through
few earthquakes originate at shallow depths
rock at speeds of about two-thirds the speed
(in the top one or two kilometres). It is
of the P wave (compressional wave), whose

common for surface rocks to be folded in
motion is in the longitudinal and thus
response to faulting at depth, giving a
strongest on the vertical component on the
monocline and scarp at the surface, but
ground surface. The S wave is a transverse
without a surface fault. In cratonic regions
wave with motion perpendicular to the
of Australia, which include most of the
direction of propagation, and is strongest in
western part of Australia, rock near the
the horizontal plane on the ground surface.
surface is quite strong and shallow
The amplitudes of the P and S waves
earthquakes are more common that in non-
decrease with distance due to geometric
cratonic regions.
spreading and absorption of wave energy.
Surface rupture occurs when a fault break
The stored compressional and shear strain
reaches the ground surface. It may produce

energy are the tectonic energy that will be

a vertical or horizontal offset, or both, with
released in the earthquake.
a displacement of millimetres to a few
The strain energy accumulates slowly over a metres, and a length from metres to tens of
kilometres. Surface rupture has occurred

long period, until it exceeds the strength at a

point on the fault plane, which then frequently in the past century in cratonic
ruptures. The rupture front spreads over the regions of Australia, but is expected to be
fault plane at about 3 km/s, with one block less common in non-cratonic regions of
moving at about 1 m/s relative to the other, Australia.
decreasing with time due to friction and loss
A site will have surface rupture potential if
of available strain energy.
an existing fault is found which has been
Seismic waves leave the fault rupture at a active during the current stress regime.
wave velocity depending on local rocks, but Clark et al. (2011) describe neotectonic
typically 4 to 6 km/s for P waves, with features have undergone displacement under
ground motion velocities starting at about the current stress regime in Australia, and
the relative block motion velocity (1 m/s). hence may have the potential for
displacement in the future. The age of the
Surface Rupture current stress regime in Australia is
estimated to lie in the range of 5 to 10
Almost all earthquakes, especially larger million years (Sandiford et al., 2004).
earthquakes, occur on existing faults. This is

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis
C2.1.1.3. Earthquake Magnitude Moment magnitude is the scale most
Earthquakes vary enormously in size. commonly used for engineering
Richter (1935) defined a magnitude scale to applications.
indicate the size of an earthquake as
These magnitude scales give similar values
M L = log10 A that can range from 0.0 to over 9.0. For
each unit of magnitude there is a tenfold
where increase in ground displacement, and a
thirtyfold increase in seismic energy release.
ML is the Richter local magnitude
scale Another measure of earthquake size is the
fault area, or the area of the fault surface,
A = maximum seismic wave which is ruptured. The fault area ruptured in
amplitude (in thousandths of a an earthquake depends on the magnitude
millimetre) recorded from a and stress drop in the earthquake. Typically,
standard seismograph at a a magnitude 4.0 earthquake ruptures a fault
distance 100 km from the area of about 1 square kilometre, magnitude
earthquake epicentre 5.0 about 10 square kilometres, and

or magnitude 6.0 about 100 square kilometres
(perhaps 10 x 10 kilometres).
M L = log10 A - F( D ) + k
There are approximate relationships
where between the magnitude of an earthquake and
the rupture area of a fault and other
F(D) = distance correction; k = parameters, such as Gibson (1994) as shown
scaling constant. in Table C2.1 and Leonard (2010).
This allows calculation of ML from
seismographs which have recorded
the earthquake at different distances

from the earthquake.

Other magnitude scales have been defined,

including the Moment Magnitude (MW)

which assigns a magnitude to the earthquake

in accordance with its seismic moment MO,
which is directly related to the energy
released by the earthquake:

M W = (log10 M O /1.5) - 10.7

MO is the seismic moment in dyn-cm.

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis
Table C2.1: Approximate measures of shallow earthquake ruptures. (Modified from Gibson,

Magnitu Rupture Typical Fault Slip Rupture Energy Global Average

de Area Length x ~Length/20,00 Duration Released Number
(km2) Width 0 ~Length/3 (MJ) per year
(km x km) (metres) (seconds)
2.0 0.01 0.1 x 0.1 0.005 0.03 60 2,000,000

3.0 0.1 0.3 x 0.3 0.015 0.1 2,000 200,000

4.0 1 1x1 0.05 0.3 60,000 20,000

5.0 10 3x3 0.15 1 2 million 2,000

6.0 100 10 x 10 0.50 3 60 million 200

7.0 1000 30 x 30 1.5 10 2,000 20
8.0 10,000 200 x 50 5.0 60 60,000 1.0

C2.1.1.4 Maximum Credible Earthquake Very few intraplate earthquakes are larger
Magnitude than magnitude 7.5. A series of very large
In view of the impracticability of identifying intraplate earthquakes in the New Madrid
all the faults on which major earthquakes area of Missouri, USA, in 1811 to 1812, had

may occur, it is necessary to use original published magnitude values

probabilistic methods to estimate expected exceeding 8.0 (Johnston and Shedlock,
ground motion versus Annual Exceedance 1992). Recent authors believe that the New
Probability (AEP). For this, it is necessary Madrid earthquakes were considerably
smaller than this (Evernden, 1975, M6.9;

first to assess the magnitude of the

maximum credible earthquake, Mmax. Gomberg, 1992 M7.3; Hough et al, 2000).

If the earthquake catalogue only covers a There is general consensus among

short period compared with the required Australian Seismologists that the maximum
return period, then the activity from a large credible magnitude of an earthquake in
surrounding area may be considered when Australia is about magnitude 7.5. This is
estimating Mmax, perhaps as large as the based on consideration of the limited depth
whole of Australia or even including other range at which crustal earthquakes can
intraplate areas over the earth. To give some occur, and the sensible maximum length of
appreciation of the likely maximum credible fault that will rupture. It was originally
magnitude, Seismologists have considered suggested that a magnitude 7.5 earthquake
the credible maximum lengths and widths of could occur anywhere in the country, but it
faults that may rupture to cause the is possible that this value may be reduced if
earthquake. Because of the limited there is evidence that active faulting,
seismogenic width of the shallow crust, large especially surface faulting, has not occurred
earthquakes have long rupture lengths, as in particular locations.
indicated in Table C2.1.

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis
Given the shallow seismogenic depths in increases, but sometimes deteriorating as
Eastern Australia, the value of Mmax 7.5 is individual seismographs are removed, or
possibly conservatively high. However for seriously deteriorating as seismograph
return periods to 1000 years, (annual networks are removed.
probability of 0.001) decreasing Mmax to
7.2 or 7.3 would give negligible effect on The coverage of a zone depends on the
probabilistic ground motion estimates. For location of seismographs relative to the
longer return period (lower annual zone. If seismographs are aligned in one
probability) motion, the lower value could direction relative to the zone, either inside or
still be justified, especially if there is little outside the zone, then coverage may vary
seismological and especially geological widely across the zone, with complete
evidence for the existence of large nearby coverage being limited to larger events.
C2.1.2.2. The Gutenberg-Richter
C2.1.2. Earthquake Recurrence Relationship

C2.1.2.1. Introduction In seismology, the Gutenberg–Richter law

Earthquake source regions are either known expresses the relationship between the

faults, or source zones where faults are magnitude and the total number of
unknown or earthquakes are distributed over earthquakes in any given region and time
many small faults. A zone can be any period of at least that magnitude.
volume in the earth, but is usually defined as
a polygonal prism with vertical sides and log = −
horizontal top and bottom. For determination
of earthquake recurrence, the main Or
characteristic of each zone is that it is
reasonable to expect that earthquake activity = 10
is uniform throughout the zone.

For known faults we need to know the type
of fault, location and dip direction. The best
· N is the number of events having a
measure of the activity on the fault is the
average slip rate, usually given in magnitude ≥

millimetres/year for active regions, or a and b are constants

metres/million years in stable regions.
The parameter b (commonly referred to as
If the earthquake catalogue included all the "b-value") is commonly close to 1.0 in
events within the entire zone, at all times seismically active regions. This means that
during the observation period, then a for given a frequency of magnitude 4.0
recurrence plot of rate of activity against events there will be 10 times as many
magnitude could be determined by simply magnitude 3.0 earthquakes and 100 times as
counting the number of events with each many magnitude 2.0 earthquakes.
magnitude and dividing by the number of
years of observation. The a-value indicates the total seismicity
rate or activity of the region.
Unfortunately earthquakes and seismograph
coverage are not so simple. Earthquakes C2.1.2.3. Random and Non-Random
cluster in time and space with long periods Processes
of quiescence. Seismograph coverage varies The earthquake recurrence rate is usually
with time, usually giving complete coverage estimated by assuming that the earthquake
to lower magnitudes as seismograph density

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis
activity rate observed over the recent past from a wide range of sources. This leads to
will continue unchanged. “large-zone hazard dilution”, where the
estimated hazard in more active regions is
Earthquake hazard studies usually assume significantly reduced, while that in larger
that earthquake recurrence is a random areas of low activity is increased slightly.
process. That is, the probability of an
earthquake does not change with space or Future earthquake hazard studies will
time, and all earthquakes are independent of continue to rely on past earthquakes, but will
each other. give greater consideration to geological and
geodetic data and processes, including stress,
One of the most obvious examples of non- strain energy, and the dynamics of
random activity in time is the foreshock- deformation in a complex geological
main shock-aftershock sequence. In hazard environment.
studies, if it is assumed that the main shock
will be the most damaging event in the C2.1.2.4. Time-Magnitude Plots
sequence, foreshocks and aftershocks can be The earthquake recurrence or seismicity
removed (de-clustered) and the recurrence of (seismic activity) of an area must take the
main shocks is determined, normally leading range of earthquake sizes into account.

to an increase in estimated hazard because There are many more small earthquakes than
the “b-value” is then lower. large. As discussed above, in most places
around the earth the b-value is 1.0. The b
Non-random activity in time is clear in an
value may be 1.3 or higher if there are many
earthquake cycle represented by a sequence small earthquakes, or 0.7 or lower if there
including a period of quiescence, possible are relatively few small earthquakes.
precursory events for large earthquakes,
foreshocks, main shock, aftershocks, The recurrence rate is complicated by the
possible adjustment events and another wide range of earthquake magnitudes that
quiescence period. are reported, or recorded and located. Most
historical earthquakes larger than about

Non-random activity in space is largely magnitude 4.0 or 5.0 are reported in

related to faults. Large earthquakes only newspapers because of the intensity of
occur on large faults and small earthquakes ground motion that was felt, or the damage
most often occur on small faults. This means that was caused. National seismograph

that the relative number of small to large networks can currently record all
earthquakes, the Gutenberg-Richter b-value, earthquakes exceeding about magnitude 2.0
varies depending on local geology, not only to 3.0 depending on the seismograph density
in its value but also in the magnitude range and distribution. A dense local network of
of applicability. Variations in b-value are seismographs about a site can record
particularly obvious with earthquake depth earthquakes exceeding about magnitude 0.0
with high values at shallow depths and low to 1.0.
values for deep earthquakes. There is
increasing evidence of earthquake activity The Gutenberg-Richter relationship is used
being affected by recent past activity in to quantify the earthquake recurrence for a
neighbouring regions, both on a single large region. The activity (a) is given by the
fault (or plate boundary) and on other nearby intercept on the vertical axis, the relative
faults. number of small to large magnitudes by the
gradient (or b-value) and a maximum
One of the most common ways to force the magnitude is applied.
seismicity of an area to fit a single
Gutenberg-Richter distribution is to use The b-value is not a constant, but varies with
large zones that include many earthquakes space and time. If a number of different

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis
processes are included, each with their own This assumption of temporally random
b-value and other parameters, the apparent earthquake occurrence clearly does not apply
seismicity of the region may still have a to earthquake clusters consisting of possible
well-defined b-value, but at some foreshocks, a main shock, and aftershocks.
intermediate value depending on the region For many years, standard practice has been
and time interval considered (the average of to consider the recurrence of clusters rather
a set of straight lines gives another straight than individual events, with the cluster
line, provided that linearity occurs for all represented by the main shock, which is the
earthquakes over the magnitude range event in the cluster that will give strongest
considered). motion, and usually (but not always) cause
the most damage. A de-clustered earthquake
If the seismicity is dominated by a swarm of history is used, and foreshocks and
shallow earthquakes then the observed b aftershocks are simply not considered in the
value will be too high to allow extrapolation earthquake recurrence calculations.
to higher magnitudes, resulting in hazard
estimates too low, or if it is dominated by a
few moderate to larger earthquakes (or the
catalogue is incomplete), the observed b-

value will be too low to allow reliable
extrapolation, resulting in hazard estimates
that are too high.

Figure C2.4: Earthquake history for the West Sydney Basin zone.
As an example the seismological history of for complete coverage of larger magnitudes,
the West Sydney Basin zone is shown in generally decreasing as the seismograph
C2.4, with the dashed line showing the network density improves over time.
estimated variation in magnitude threshold

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis
The difference in rates of activity for show clusters, usually a main shock with
magnitudes 1.5 to 2.2 between 1970 and foreshocks and aftershocks.
1988, and between 1992 to the present is
probably because the original catalogue was The next stage in quantifying the activity
not fully “de-blasted”, so contains many within the zone is to convert the historical
quarry and mine blasts as well as data into cumulative recurrence data. This is
earthquakes. The vertical lines of events shown in Figures C2.5 and C2.6.

Figure C2.5: Earthquake magnitude recurrence, not de-clustered.

Figure C2.6: Earthquake magnitude recurrence, de-clustered.

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis
Note that in this example that the earthquake greater earthquake activity rates than the
with a 1/1000-year average recurrence extrapolated seismicity estimates (Figures
interval is magnitude 6.0 for the original C2.6 and C2.7,). Earthquake recurrence
data and 6.4 for the de-clustered data. The estimates made by extrapolating small
corresponding values for 10,000-year earthquake recurrence will under-estimate
average recurrence interval are magnitude the hazard in these regions.
6.9 and magnitude 7.2 respectively.

C2.1.2.5. Geologically Observed Slip

Rates vs. Extrapolated Observed
Larger active faults appear to have
geologically observed slip rates that require


Figure C2.7: Earthquake magnitude recurrence for the Newcastle zone.

In areas with no obvious large active faults, will over-estimate the hazard in these
such as large areas of undeformed flat-lying regions.
sediments, the observed recurrence rates are
lower than the extrapolated estimates (or C2.1.3. Seismic Ground Motion
zero), with fewer (or not so large)
C2.1.3.1. Ground Motion Measures
earthquakes. This suggests that these smaller
earthquakes occur in regions with small Earthquake ground vibration is recorded by
faults, and lower maximum magnitude. a seismograph or a seismogram. Most
Earthquake recurrence estimates made by modern seismographs record three
extrapolating small earthquake recurrence
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis
components of motion: east-west, north- The larger the earthquake magnitude, the
south and vertical. greater the amplitude, the longer the duration
of motion, and the greater the proportion of
The rupture time for a small earthquake is a seismic energy at lower frequencies. A small
fraction of a second, for earthquakes of earthquake has low amplitude (unless it is
magnitude 5.0 it is about a second, and for very close), short duration, and has mainly
large earthquakes may be up to tens of high frequencies.
seconds. However the radiated seismic
waves travel at different velocities, and are The smaller the distance from an earthquake
reflected and refracted over many travel to the site, the higher the amplitude. The
paths, so the total duration of vibrations at a duration is not strongly affected by distance.
site persist longer than the rupture time, and High frequencies are attenuated by
show an exponential decay. absorption within the ground more quickly
than low frequencies, so at greater distances
As discussed above, several types of seismic the proportion of seismic vibration energy at
wave are radiated from an earthquake. Body high frequencies will decrease.
waves travel in three dimensions through the
earth, while surface waves travel over the C2.1.3.2. Earthquake Intensity

two-dimensional surface like ripples on a Earthquake Intensity is a measure of the
pond. There are two types of body wave (P effect of the seismic waves at the surface,
waves and S waves), and two types of and is normally given on the Modified
surface waves (Rayleigh and Love waves). Mercalli Intensity scale; a copy of which is
attached in Appendix A. This is an arbitrary
Primary or P waves are ordinary sound scale defined by the effects observed
waves travelling through the earth. They are (whether sleeping people were woken, trees
compressional waves, with particle motion shaken, etc.) and on the damage caused.
parallel to the direction of propagation. Normally the maximum intensity occurs
Secondary or S waves are shear waves, with near the epicentre of the earthquake, and
intensity then decreases with distance.

particle motion perpendicular to the

direction of propagation. The amplitude of S However this may be affected by the
waves from an earthquake is usually larger orientation of the rupture, or local ground
than that of the P waves. conditions such as topography or surface

P waves travel through rock faster than S

waves, so they always arrive at a C2.1.3.3. Attenuation and Amplification
seismograph before the S wave. of Ground Motion
Seismic ground motion generally attenuates
The frequency content of seismic ground with increasing distance from the source due
motion covers a wide range of frequencies to radiation and hysteretic damping. High
up to a few tens of hertz (cycles per second). frequency motion is attenuated more quickly
Most engineering studies consider motion with distance than lower frequency motion.
between about 0.2 and 25 Hz.
In the old, hard rocks in cratonic regions
The amplitude, duration and frequency such as the shield areas in Western Australia
content of seismic ground motion at a site and in eastern Canada, earthquakes of
depend on many factors, including the magnitude 5.0 can be felt at much greater
magnitude of the earthquake, the distance distances than in Eastern Australia, over 400
from the earthquake to the site, and local site km compared with less than 150 km, and
conditions. cratonic ground motion models must reflect
this. Unfortunately, there is very little strong
motion data available from such areas.
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis
In cratonic areas there are often very hard their mass and stiffness, usually in the range
crystalline rocks to near the surface and from about 0.5 to 5 Hz for embankment
earthquake depths tend to be quite shallow. dams and 2 hertz to 20 hertz for concrete
This can lead to generation of strong surface gravity dams.
waves by relatively small earthquakes with
long period motion (about 2.0 seconds in the C2.1.3.4. Site Response – Near-Surface
Yilgarn of Western Australia). Ground Amplification and Attenuation
motion models for this region (Liang et al., Ground motion at a site on the Earth’s
2008; Somerville et al., 2009) incorporate surface may be significantly affected by
such unusual characteristics. near-surface geology and topography. There
are many phenomena that affect near-surface
In selecting attenuation relationships or motion, including:
ground motion prediction equations care is
needed, and attention paid to the mechanism · Impedance amplification
of the source earthquake, e.g. whether · Resonance of surface sediments
shallow intraplate or deep plate boundary · Additional sedimentary basin effects
earthquakes. · Variation of frequency-dependent

attenuation, Q(f), in surface sediments
At low amplitude levels, the ground motion
is amplified by the near-surface geology · Scattering in complex geology
because it is generally more flexible than · Reflections in stratified surface geology
· Groundwater
unweathered bedrock. The amplification
increases as the shear wave velocity of the · Wave conversions at interfaces
surface geology decreases due to impedance · Focusing by complex structures
amplification. The shallow shear wave · Topographic effects
velocity is quantified, for example, by Vs30,
the time-averaged shear wave velocity in the Combinations of impedance amplification
upper 30 metres of the ground. However, at and resonance can give amplifications of
high amplitude levels in soft soils, weak bedrock ground motion that can exceed x 4

surface materials absorb seismic energy to x 8 at resonant frequencies. Resonant

rather than transmit it unchanged, thus motion gives greater amplification to
tending to reduce amplitudes at the surface. horizontal motion than to vertical motion.
The amount of attenuation depends on the Site response can vary over very short

properties of the materials, and especially distances, such as tens of metres.

their thickness. Site response can be measured by comparing
In horizontally stratified sediments, the near surface motion with bedrock motion.
surface layers will vibrate preferentially at Measurement of ambient vibrations at a
their own natural frequencies, depending on point on the surface can clearly indicate the
their thickness and elastic properties. The natural frequencies of the sedimentary
earthquake motion at the natural frequencies column simply because the spectrum shows
of the near-surface layers is amplified, while a high peak. In other cases using single point
motion at other frequencies may be little measurements, the ratio of horizontal
affected or even attenuated. The spectral motion over vertical spectral motion
amplification effect can be especially will give some idea of both natural
pronounced for deep soft sediments that frequencies and the degree of amplification,
have very low damping, such as those because horizontal motion is amplified much
underlying Mexico City. more than vertical motion.
Dams (like all other structures) each have Measuring ambient vibrations with a small
their own natural frequency depending on array can reveal more information about the
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis
sedimentary column, such as the Spectral resonance has amplified motion at the low
Autocorrelation (SPAC) method which gives resonant frequency.
a velocity dispersion curve from which the
variation of shear wave velocity with depth Local site conditions on a relatively small
can be estimated. These methods usually scale, particularly soft surface sediments but
consider depth ranges from metres to also topography, can significantly affect
hundreds of metres. surface motion from an earthquake. These
effects can vary rapidly with location,
Site amplification has little effect on hard showing significant changes over distances
rock sites, but can dominate the hazard on within tens or hundreds of metres.
sites with deep soft sediments.
Estimating Site Response in Practice
Free-Surface Amplification
Site response, or site amplification, depends
The amplitude of seismic waves increases as on many phenomena, including variation in
they approach the Earth’s (free) surface, at impedance depending on the rock properties,
which they are reflected. This amplification variations in attenuation especially through
can be considered in terms of constructive soft rocks, resonance in surface layers

interference between the up going and down especially in soft rock, resonance of
going waves, and can give surface motion up sedimentary basins, conversion of wave
to about double that at depth for a uniform types between P, S and surface waves at the
half space. However, as noted above, the surface or at boundaries beneath the surface,
near surface material usually has lower shear and focusing of waves by irregular surfaces.
wave velocities than bedrock, so the surface
motion is usually more than twice that at The treatment of each of these can be from
depth. The nature and degree of free surface simplistic to complex. The critical
amplification varies with topography, even phenomena will vary from site to site. Some
in fresh, strong rock. Changes in soil of the phenomena are period (or frequency)
thickness above an irregular bedrock surface dependent and require spectral variations

can give complex surface amplification that (e.g. resonance, attenuation), and others vary
varies with wave duration. little with frequency (e.g. impedance
For many earthquake hazard studies it is

Resonance in the surface sediments causes normal practice to calculate bedrock motion
amplification at particular frequencies, for surface outcrop or at the surface of
especially at the natural frequency of the bedrock below sediments. Unless the site is
sediments. This depends on the thickness near to an active fault, there is little variation
and elastic properties of the sediments. in estimated bedrock motion over distances
Seismic motion recorded on hard rock tends of kilometres, and the same bedrock motion
to be broadband, while that recorded on soft can be applied at various locations about the
sediments is usually dominated by the sites. That does not hold if the rock
resonant frequency. foundation conditions are greatly different
between the river bed and the abutments.
Attenuation The bedrock motion can then be used to
In surface sediments, high frequency estimate surface motion or motion of the
vibrations are attenuated much more with dam, listed in order of increasing
distance than low frequencies. If sediments complexity, accuracy and cost:
are very thick, much of the high frequency
motion will be lost and peak surface
accelerations will be relatively low, even if
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis
· When using Australian Loading code · If site specific response must be
AS1170.4 (2007), the bedrock ground considered, and if the site has near
motion values given would be multiplied horizontal stratification, the results
using the AS1170.4 site sub-soil class provided can be used to select design
factors, which vary from 0.8 for hard earthquakes for use with a SHAKE type
rock to 3.5 for very soft soils. This does program to estimate surface motion and
not consider frequency dependent cyclic stresses in the foundation
resonance. AS1170.4 has limited overburden and embankment. The site
application for dams because it only is represented by a one-dimensional
provides loads for the 10% chance of model giving the shear wave velocities
exceedance in 50 years, or 1 in 475 of each layer down to bedrock (not just
annual probability. the surface 30 metres), plus densities,
· For the next level of design, for and the depths of the layer interfaces. If
horizontally stratified surface layers, a conditions are appropriate and the shear
local value of Vs30 allowing for the Vs wave velocity variation is known or can
of the sediments overlying bedrock can be estimated, this is relatively easy to
be used to recalculate the results to give do.

an average estimate of site response that · If frequency dependent site response is
considers impedance amplification, but considered, and if the site has
does not consider frequency dependent topographic variations, or does not have
phenomena such as resonance. Vs30 is near horizontal sub-surface
the time averaged shear wave velocity in stratification, the results provided can
the top 30 metres of rock and sediments, be used to select design earthquakes for
calculated from the inverse of the mean use with a finite-element or finite-
slowness in the top 30 metres (using difference computer program to
seconds per metre rather than metres per estimate surface motion. The site is
second), or 30 divided by the total represented by a two- or three-
vertical shear wave velocity (Vs) travel dimensional model giving the

time through these 30 metres. This gives distribution of sub-surface materials. In

a greater weighting to low velocities some cases this may be a combined site-
rather than to high velocities. Use of structure model, incorporating site-
Vs30 to estimate soil amplification at a structure interaction. Two-dimensional

specific site assumes that the average models require much data and complex
shear wave velocity, shear modulus computation, and three-dimensional
reduction, and damping increase profiles models require a great deal more. This
represented in the strong motion is seldom done in practice as one
database used to develop the ground dimensional models are usually
motion prediction equations are a adequate. More than one 1D model may
reasonably accurate representation of the be used if conditions vary across a dam
specific profiles at the site. If this is not site.
the case, then a site-specific analysis of · A time-consuming and generally
the soil amplification (described next) impractical method in anything other
should be done. than highly seismically active areas
is to determine a site transfer
function empirically by installing
seismographs at the site and nearby
on bedrock (either in a borehole or
on a rock outcrop within a distance
of some kilometres). Comparison of
spectra from regional and distant
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis
earthquakes will give frequency normally felt rather than heard, with any
dependent site response for short and sounds being the reaction of a building to the
long period motion respectively. vibrations rather than the earthquake itself.
A very small proportion of these felt Papua
In the case of foundations on rock with a New Guinea earthquakes, perhaps about
fairly typical Vs30 of 760 m/s, there will be 0.1%, will cause damage in their epicentral
relatively little resonance or other frequency area, and earthquakes smaller than
dependent site response. Inclusion of the magnitude 6.0 rarely cause damage.
Vs30 term will account for some impedance
amplification. There are a number of factors that influence
the estimation of earthquake hazard in
Australia. One is the short duration of
C2.1.4. Earthquake Hazard in Australia
documented history, with a little over 200
C2.1.4.1. Australian Earthquakes years of data about Sydney, and
The Australian continent is within a tectonic considerably less for most of the continent.
plate shared with most of India and some of Another factor is the large area of the
New Zealand, and all of its earthquakes are country relative to the size of the population.

intraplate. The plate boundaries to the north Seismographs are distributed relatively
and east are among the most active on the sparsely, limiting the accuracy of earthquake
Earth. Possibly as a result of this, Australia locations and magnitudes.
is one of the most active intraplate areas on
the Earth. Despite this, the hazard is quite As was the case over most of the Earth,
low when compared with active plate seismograph coverage of local earthquakes
boundary areas. in Australia was only established in about
1960, following the International
Most people in Australia can expect to feel Geophysical Year. While there should be
an earthquake about every five to ten years, some link between population and
although many of these will not be seismograph density with more

recognised as an earthquake. Most instrumentation in the populated southeast,

Australian earthquakes that are reported are coverage is highly variable and still far from
smaller than magnitude 3.0 and are heard complete. The Australian National
with a noise like a distant quarry blast or Seismograph Network, operated by
thunder (due to coupling of P-waves in the

Geoscience Australia, aims to locate all

ground into sound waves in the air), often earthquakes in Australia larger than
followed by a slight vibration. magnitude ML 3.0. Local seismograph
Only a small proportion of the earthquakes networks have non-uniform coverage, often
that are felt, perhaps about 5%, will cause with good coverage of large dams and poor
any damage in their epicentral area. If they coverage of major cities.
occur in an inhabited area, most earthquakes Figure C2.8 shows the location of Australian
of magnitude 4.0 to 5.0 will cause some earthquakes with magnitude exceeding ML
minor damage, while magnitudes greater 4.0 in the period 1850 to 2014. Note that
than 5.0 may cause significant damage. many earthquakes prior to 1960 were not
By contrast, in an active plate boundary area located so are not shown.
like New Britain or Bougainville in Papua
New Guinea, earthquakes are felt very often,
on average every week or two. These are

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis

Figure C2.8: Australian earthquakes exceeding magnitude 4.0 from 1850 to 2016
Table C2.2 lists some of the largest earthquakes which have occurred in Australia. It can be seen

that there are several in the range M6.5 to 7.0.


ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March
2017 61
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis
Table C2.2: Significant Australian earthquakes

Date Place Mag Imax Damage, Approximate A$(2016).

1892- Flinders Island, MP 7+ Offshore epicentre, little damage. Felt through
01-26 Tas 6.9 Tasmania and East Victoria.
1897- Beachport, SA MP 8 Damage in Kingston, Robe, Beachport, major
05-10 6.7 liquefaction; felt from Ceduna to Melbourne.
1902- Warooka, SA MP 8 At Warooka chimneys fell, walls partially
09-19 6.0 demolished, few buildings without damage.
1903- Warrnambool, MP 7 Extensive minor damage and liquefaction at
07-14 Vic 5.3 Warrnambool. Followed similar event in April.
1918- Bundaberg, Qld MP 6 Epicentre offshore, minor damage in
06-06 6.0 Rockhampton.
1941- Meeberrie, WA MS 8 Isolated area. Damage to remote farm houses
04-29 7.2 including cracked walls, burst tanks.

1954- Adelaide, SA ML 8 Widespread minor damage, A$200m. Epicentral
02-28 5.4 was rural but is now urban.
1961- Robertson, ML 7 Damage in the Moss Vale, Robertson and Bowral
05-21 NSW 5.6 area, A$10m.
1968- Meckering, WA MS 9 Most buildings in Meckering destroyed, $90m.
10-14 6.8 32 km surface rupture.
1969- Boolarra, Vic ML 6 Cracked walls and fallen chimneys in the
10-14 5.6 epicentral area.
1973- Burragorang, ML 6 Minor damage in Picton, Bowral and

03-09 NSW 5.0 Wollongong, A$7m.

1979- Cadoux, WA MS 9 Many buildings at Cadoux destroyed, $9m. Only
06-02 6.2 one injury. 15 km surface rupture.
1986- Marryat Creek, MS 8 Epicentre in remote area. Cracked walls in

03-30 SA 5.9 nearest homestead, 13 km surface rupture.

1988- Tennant Creek, MS 9 Epicentre in remote area. Damage A$4m, mainly
01-22 NT 6.8 to damaged gas pipeline, 35 km surface rupture.
1989- Uluru, NT mb 6 Minor damage at Uluru National Park (Ayers
05-28 5.7 Rock). Epicentre west of Mt Olga.
1989- Newcastle, ML 8 Thirteen fatalities, $4500m plus damage.
12-27 NSW 5.6 Widespread minor to moderate damage.

C2.1.4.2. Mechanism of Australian failure is on a fault that is oriented at an

Earthquakes angle other than 90° to the principal stress
Almost all Australian earthquakes have direction.
mechanisms with the maximum principal Compression giving reverse and thrust faults
stress near horizontal. Most have the produces surface uplift, so these earthquakes
minimum principal stress near vertical, are most likely to occur in areas where
producing reverse or thrust faults. There may mountains are developing (such as the
be some strike-slip movement, when the
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis
eastern Highlands or Flinders Ranges), and of their relatively shallow depths and the low
less likely under sedimentary basins (such as density of seismographs.
the Murray-Darling Basin or the Great
Artesian Basin). If it is not possible to constrain an
earthquake depth using seismograph data or
There are a number of active faults upon other observations (e.g. a magnitude 1.0
which several episodes of movement have earthquake that is felt must be within a
been proven in the last 100,000 years. These couple of kilometres of the surface), an
include the Edgar fault; Tasmania, Hyden arbitrary “normal” depth may be used.
fault, Western Australia; Wilkatana and Typical values used in Australia include 0, 5,
Roopena, South Australia; and Cadell, 10 or 33 kilometres. These may or may not
Victoria/New South Wales (Clark et al, be realistic. Some observatories may select a
2012). depth from a range of standard values,
depending on the character of the recorded
Large Australian earthquakes (e.g. Tennant waveforms (e.g. a small magnitude
Creek, 1988), have occurred on faults had earthquake with large surface waves must be
not been previously been recognised and shallow).
mapped. Crone et al. (1997) describe

evidence that there had been no movement Eastern Australian earthquakes are usually at
on the fault at Tennant Creek for 200,000 depths between 1 and 20 km, while those in
years. Central Australia may be a little deeper, and
those in Western Australia may be a little
Based on the available information, it would shallower. Australia’s deepest known
appear that differences between earthquake earthquake occurred at 39 km depth offshore
ground motion s in non-cratonic regions of from Arnhem Land in 1992 (McCue and
Australia and those from reverse fault Michael-Leiba, 1993).
earthquakes in USA, China, etc., data from
which are included in some empirical In Eastern Australia, earthquakes at depths
ground motion prediction models are based, of less than 5 km are regarded as shallow,

are not sufficiently great as to invalidate the and greater than 15 km as deep. The
use of those models in Australia. This is an Newcastle earthquake of December 1989
aspect that will need to be further assessed was at a depth of about 12 km.
as more ground motion data from Australian

earthquakes is gathered. All of these Australian earthquakes are very

shallow compared with those in plate
More detailed design methods are based boundary areas, where subduction
upon seismic response analyses to actual earthquakes can occur as deep as 700 km,
ground motions rather than empirical and depths of less than 70 km are regarded
methods so this is not an issue provided as shallow, and greater than 300 km as deep.
there is a careful selection of ground motion
histories to represent the earthquake It seems that small earthquakes tend to occur
conditions in Australia. more often at shallow depths, while
moderate to large earthquakes rupture at
C2.1.4.3. Earthquake Depths greater depths (Allen, 2004). Many of
Usually, earthquake depths can be precisely the larger Australian earthquakes have
determined from local networks only if the occurred in the cratonic regions and have
distance to the nearest seismograph is not rupture through to the surface due to the
greater than about twice the earthquake high strength of the surface rocks, giving a
depth. The depths of most Australian surface fault rupture.
earthquakes are poorly constrained, because
Shallow earthquakes often have many
aftershocks, some with magnitudes
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis
approaching that of the main shock. Deep · Low maximum magnitude, perhaps Mw
earthquakes usually have few aftershocks, 7.5
and these are usually no larger than one · Hazard is quite widely distributed
magnitude unit below the main shock
magnitude. Shallow, from surface to 20 km

Because of their shallow depth, small · Small events often felt and heard within
Australian earthquakes are often heard or a very limited area
felt. Magnitude 1.0 events can be felt to a · Moderate magnitudes can cause damage
distance of about 1 km, and magnitude 2.0 to · Above Mw 6 usually gives surface
about four kilometres. These are slant rupture, especially in cratonic regions
distances, so only very shallow events of
these magnitudes will be felt. Dominant horizontal compression

Because the short travel distances from · Reverse faults predominate, strong
shallow earthquakes do not give much and relatively inactive, so stress
attenuation of high frequency vibrations, and levels are high
if a relatively high stress drop gives a high · Earthquakes may have high stress

proportion of high frequency vibration from drops, giving high frequency, high
the earthquake source, these events have acceleration, short duration motion
enough energy in the audio range to allow
C2.1.4.4. Australian Earthquake Hazard
them to be heard. For many such small
earthquakes the sound heard is more Maps
significant to an observer than the vibrations As presented in ANCOLD (1998)
felt. earthquake hazard maps for Australia have
changed considerably over the years as more
For similar reasons, moderate magnitude earthquakes are recorded in an increasing
earthquakes in Australia produce motion seismograph recording network, and as
with strong peak ground accelerations, and earthquakes such as Tennant Creek have

can cause significant damage. The occurred in what was thought to be a low
Newcastle earthquake was only of seismic hazard area.
magnitude ML 5.6, but caused extensive
damage. However much of this occurred in In practice these hazard maps are of limited

areas where the significant depth of alluvium use for dams because they only consider the
overlying bedrock amplified the ground 10% chance of exceedance in 50 years, or 1
motions. in 475 / annum hazard which is too frequent
for design ground motions for most dams.
In summary, Australian earthquakes are:
As a result site specific hazard studies are
Intraplate and continental, so they are generally required for dams.
C2.1.5. Reservoir Triggered Earthquake
· Most people feel earthquakes just a few Reservoirs may induce seismicity by two
times in their lifetime mechanisms. Either the weight of the water
· However, Australia is one of the most may change the stress field under the
active intraplate areas reservoir, or increased ground water pore
pressure may decrease the stress required to
Distributed over many small to moderate cause an earthquake. In either case, reservoir
faults triggered earthquakes (RTE) will only occur
if relatively high stresses already exist in the
· Fault lengths rarely exceed 100 km,
area. If a recent large earthquake has
relieved the stress, perhaps in the last few
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis
hundred years for low seismicity areas like and small shallow induced events have
Australia, then RTEs are unlikely to occur. probably occurred under many others. A
relatively high proportion of reservoirs with
RTE events initially occur at shallow depth RTE seismograph networks do record such
under or immediately alongside a reservoir. activity. A high proportion of RTE examples
As years pass after first filling, and occur in intraplate areas, with above average
groundwater pore pressure increases rates in China, Australia, Africa and India.
permeate to greater depths and distances, the
events may occur further from the reservoir. It is not easy to predict whether a particular
This occurs at a rate of something like one new reservoir will experience RTE because
kilometre per year. RTE’s are experienced the stress and strength at earthquake depths
under new reservoirs, usually starting within cannot normally be measured. For the same
a few months or years of commencement of reason, prediction of normal tectonic
filling, and usually not lasting for more than earthquakes has been unsuccessful in most
about twenty years. Once the stress field and parts of the world.
the pore pressure field under a reservoir have
stabilised, then the probability of future It seems that RTE with many small events is
earthquakes reverts to values similar to those more likely to occur in intraplate areas with

estimated if the reservoir had not been near-surface crystalline rocks like granite,
produced. Most of the earthquake energy rather than sedimentary rocks. A larger
does not come from the reservoir, but from magnitude RTE event can only occur if there
the normal tectonic processes. The reservoir is an existing fault of sufficient dimension
simply acts as a trigger. that is late in its earthquake cycle, with the
stress already approaching the strength of
If there is a major fault near the reservoir, the fault.
RTE events can exceed magnitude 6.0 (e.g.
Xinfengjiang, China, 1962, M 6.1; Koyna, ICOLD (1983, 1989, 2012), conclude that
India, 1967 M 6.3). Such events will only there is documented evidence to prove that
occur if the fault is already under high stress. impounding of a reservoir sometimes results

Several Australian reservoirs may have in an increase of earthquake activity at or

triggered earthquakes exceeding magnitude near the reservoir. ICOLD (1983) conclude
5.0 (e.g. Eucumbene, 1959, M 5.0; that:
Warragamba, 1973, M 5.0; Thomson, 1996,

– Earthquakes of magnitude 5 to 6.5

M 5.2), and others may have triggered
were induced in 11 of 64 recorded
smaller earthquakes.
It is more common for a reservoir to trigger – The greatest seismic events have been
a large number of small shallow associated with very large reservoirs
earthquakes, especially if the underlying (but there is insufficient data to show
rock consists of jointed crystalline rock like any definite correlation between
granite (e.g. Talbingo 1973 to 1975; reservoir size and depth and seismic
Thomson 1986 to 1995). These events activity).
possibly occur on joints rather than – In view of the above, a study of
established faults, so are limited in size, and possible induced seismic activity
only give magnitudes up to 3 or 4. Their should be made at least in cases
shallow depth means that even events where the reservoir exceeds 5x108 m3
smaller than magnitude 1.0 may be felt or in volume, or 100 m in depth.
heard. – The load of the reservoir is not the
significant factor; rather it is the
RTEs have been observed for over one increased pore water pressure in
hundred reservoirs throughout the world,

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis
faults, leading to a reduction in shear assumed to be potentially active in assessing
strength over already stressed faults. the PSHA. They represent faults which have
ICOLD (2012) indicate that there are so far ruptured in the last 5 to 10 million years.
only six generally accepted cases of C2.3 General Principles for Selection of
reservoir triggered seismicity where the Design Ground Motion and Analysis
magnitude of the event exceeded 5.7. The Method
largest recorded magnitude event that is There are two ways to select the design
believed to be due to a reservoir-triggered ground motion. These are related to the
event is 6.3. analysis method. They are:

C2.2Terminology (a) Deterministic Analysis. Design ground

Active fault. The definition adopted for an motions at the dam site are defined in
active fault is adapted from ICOLD (2016a). probabilistic terms; and where there are
active faults in the vicinity of the dam,
ICOLD (2016a) include “or very long faults also by the ground motion resulting from
which have moved repeatedly in Quaternary the MCE on the faults.
(b) Risk based Analysis. A risk based

time (1.8 million years). For this guideline
this is not included as such very infrequent approach requires assessment of the
activity would be too conservative for use in seismic ground motions at the dam site
the deterministic analysis approach. These up to and beyond the SEE for use in risk
faults are included in PHSA as neotectonic analyses.
The steps involved in the risk analysis
The United States Nuclear Regulatory process are:
Commission (USNRC) define a capable
fault as a fault which has exhibited one or (i) Determine the AEP of earthquake
more of the following characteristics: ground motion (PE) over the range of
earthquake events which may affect

(a) Movement at or near the ground surface the dam. Table C2.3 gives an
at least once within the past 35,000 years example for AEP vs ground
or movement of a recurring nature within acceleration.
the past 500,000 years. (ii) Determine the conditional probability

(b) Macro-seismicity instrumentally (PBC) that for each of the ground

determined with records of sufficient motion ranges (e.g. 0.125 g to 0.175
precision to demonstrate a direct g in Table C2.3) the dam will fail
relationship with the fault. resulting in uncontrolled release of
(c) A structural relationship to an active the reservoir. In assessing this
fault according to characteristics (a) or conditional probability all modes of
(b) that movement on one could be failure should be considered and the
reasonably expected to be accompanied probabilities combined, making
by movement on the other. allowance for interdependence and
mutual exclusivity or otherwise (e.g.
This is broadly consistent with the ICOLD for embankment dams, slope
(2016) definition. Active faults are included instability, piping,
in the Probabilistic Seismic Hazard liquefaction/instability, and for
Assessment (PSHA) and in assessing the concrete gravity dams, overturning,
MCE for Standards based design. and sliding).
(iii)Assess the probability of failure for
Neotectonic fault. The definition is taken
each range of ground motion by
from Clark et al (2012). Such faults are

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis
multiplying the AEP with PBC i.e. PB (v) Assess whether the resulting risks are
= PE x PBC – see Table C2.3. tolerable.
(iv)Sum the probabilities to give the
overall annual probability of failure
due to earthquake.
Table C2.3. Example of assessing the probability of failure by earthquake.
Acceleration Annual Conditional(1) PB(2)
Probability (PE) Probability
<0.075g 0.874 0.0005 0.0004
0.075g to 0.125g 0.100 0.005 0.0005
0.125g to 0.175g 0.015 0.05 0.0007
0.175g to 0.225g 0.007 0.1 0.0007
0.225g to 0.3g 0.003 0.3 0.0009
>0.3g 0.001 0.5 0.0005
Total 1.000 0.0037

(1) Given the earthquake occurs
(2) PB = PE x PBC
Preferred method. passing the deterministic method but not
satisfying the PRA tolerable risk criteria.
The risk based approach is preferred for
assessment of existing dams for the C2.4 Description of a Probabilistic
following reasons: Seismic Hazard Assessment
1. In the deterministic approach it is C2.4.1 Inputs into PSHA.
difficult to account for the fact that the

conditional probability of failure of the (a) Earthquake Forecast (EQF)

dam given the SEE is not 1, and may be
very different for different types of dams The PSHA uses an earthquake forecast,
and foundation conditions. Hence either which predicts the locations,
magnitudes, and frequencies of

these conditional probabilities have to be

estimated, or it has to be understood that occurrence of all earthquakes that
the different dams have quite different contribute seismic hazard at the site.
annual probabilities of failure, even The earthquake forecast is based on
though they are assessed for the same earthquake source models, which are of
SEE. two kinds: distributed earthquake
2. Deterministic methods potentially mask sources, and fault sources. To account
the fact that many dams have some for epistemic uncertainty in earthquake
likelihood of failure at ground motions source models, alternative viable
less than the SEE. earthquake source models should be
3. Some design methods are by their nature used in a logic tree framework, with
probabilistic; e.g. liquefaction weights given to each alternative model.
assessments, and hence inherently risk (b) Ground Motion Prediction Equations
4. Risk based methods are commonly used
for Portfolio Risk Assessments (PRA) in Ground motion prediction equations
Australia, and if deterministic methods provide estimates of the ground motions
are used they may result in a dam at a site having specified site conditions
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis
located at a specified distance from an ground motion parameter of interest
earthquake of specified magnitude and exceeds the value ‘s’.
other source properties (such as style of Some prefer to report this in terms of
faulting). They specify the ground computing s for a certain probability of
motion by peak ground acceleration and occurrence, P, in a time period t. For
response spectral acceleration (5% example a ground motion level having a
10 -4 / annum probability of exceedance
In the past, site conditions were may be expressed as equivalent to a
parameterized in GMPE’s as a generic return period of 10,000 years.
geological category such as rock or soil, The annual probability of exceedance is
but most current GMPE’s typically now calculated as follows:
use the time-averaged shear-wave
velocity in the upper 30 metres of the
bedrock profile (Vs30) that underlies the Faults

site. f (s) = å ( òò f (m)(P(SA > s | m, r)P(r | m)dmdr) i

i =1 m,r
In cases where soil overlies bedrock, Equation 2

especially where the soils are expected
to have significantly non-linear
behavior, non-linear site response f(mi) = probability
analysis should be carried out using density function for events of magnitude
ground motions estimated at an mi
“engineering bedrock” level below the (from the EQF).
soil profile by the PSHA. These analyses
P(SA>s|m,r) = probability that
are carried out by the dam Consultant. SA exceeds s for a given magnitude m
To account for epistemic uncertainty in and
earthquake ground motion models, distance r (from the GMPE).

alternative viable earthquake source P(r|m) = probability that

models should be used in a logic tree the source to site
framework, with weights given to each distance is r,
alternative model. given a

source of
C2.4.2 The PSHA Process magnitude m
(a) Method (from the EQF).
The probabilistic analysis is performed
PSHA is based on methodology using a computer program of the type
originally proposed by Cornell (1968), described further below. The computer
outlined schematically in Figure C2.9. program should allow the treatment of
If seismicity is considered to follow a two types of uncertainty: epistemic
random Poisson process, then the uncertainty and aleatory variability.
probability that a ground motion, such as
Spectral Acceleration (SA) exceeds a (b) Treatment of Random Variability within
certain value (s) in a time period t is the PSHA Hazard Integral
given by:
Aleatory variability (randomness) results
- f (s )t
P(SA > s) = 1 - e Equation 1 from randomness in natural physical
processes that coexist in nature. The
where f (s) is the annual mean number size, location, and occurrence time of the
of events (also known as “annual next earthquake on a fault, and the site-
probability of exceedance”) in which the
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis
to-site random variability in the ground even with the collection of additional
motion level at a given distance from a data. Integration over aleatory
given earthquake (shown in Box 4 of variabilities is carried out within the
Figure C2.9) are examples of quantities hazard curve (Equation 2) to yield a
considered aleatory. In current practice, single hazard curve.
these quantities cannot be predicted,


Figure C2.9 Schematic diagram of probabilistic seismic hazard analysis. The attenuation
equations shown in Box 4 are also called ground motion prediction equations.(Source:
Paul Somerville)
(c) Treatment of Epistemic Uncertainty Epistemic uncertainty results from
outside the PSHA Hazard Integral imperfect knowledge about the process
of earthquake generation (e.g. in
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis
alternative viable earthquake forecast multiple interpretations are each
models) and the assessment of their assigned a weight within the logic tree
effects (e.g. in alternative viable ground framework, resulting in a suite of hazard
motion prediction models). Viable curves and their associated weights.
alternatives are mutually exclusive, and Examples of logic trees for distributed
are treated using logic trees outside the earthquake sources and fault sources are
PSHA hazard integral. Epistemic shown in the upper and lower parts of
uncertainties are thus expressed by Figure C2.10.
incorporating multiple hypotheses,
models, or parameter values. These


Figure C2.10. Example logic trees for the treatment of epistemic uncertainty in distributed
earthquake sources (top) and fault sources (bottom) in PSHA. (Source: Paul Somerville)

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis

C2.4.3 Products of PSHA and their C2.11 shows peak ground acceleration
analysis. hazard curves for a group of distributed
earthquake sources and fault sources,
(a) Hazard Curves shown by the colored lines, with the black
curve showing the combined total hazard.
The PSHA estimates the ground motion
level as a function of annual probability of
exceedance, or return period. Figure


Figure C2.11 Example hazard curve for a group of distributed earthquake sources and fault
sources, with the uppermost black curve showing the combined total hazard. (Source: Paul
(b) Uniform Hazard Response Spectra curves for each of a set of ground
(UHS) motion periods, including PGA, which is
equivalent to the zero period response
Uniform hazard curves for a set of return spectral acceleration (Figure C2.12).
periods are obtained from the hazard

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March
2017 71
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis

Figure C2.12. Equal hazard response spectra for a range of return periods.
(c) Fractiles of the Uniform Hazard seismic hazard curves are obtained. The
Response Spectrum (UHS) usual practice in seismic hazard analysis
is to use the mean of these hazard
There is epistemic uncertainty in the true curves, to ensure that large hazard
value of the seismic hazard results due values are given appropriate weight.

to uncertainty in the true state of nature.

In the example shown in Figure C2.13, The fractiles of these multiple hazard
three distributed earthquake source curves are used to quantify the epistemic
models (AUS5, RF and GA) were used uncertainty in the true value of the
hazard. Figure C2.13 shows the 95th,

because of uncertainty about the degree

to which each is correct. Similarly, six 85th, 50th (i.e. median), 15th, and 5th
ground motion models were used percentiles of the hazard for a given
because of uncertainty about the degree return period, as well as the mean
to which each is correct. This kind of hazard. The mean hazard is generally
uncertainty is modeled using logic trees, slightly higher than the median. The
with the alternative branches fractiles represent the degree of certainty
(corresponding to the alternative that the true value of the hazard does not
choices) being given weights. The exceed the value given by the fractile.
seismic hazard is calculated for each of For example, there is a 50% chance that
these branches, in this case, for each of the true value of the hazard exceeds the
the eighteen combinations of distributed median (50th fractile) value, and a 15%
earthquake source model and ground chance that it exceeds the 85th fractile.
motion model. As a result, eighteen

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis

Figure C2.13. Fractiles of the hazard for a return period of 10,000 years (10 -4 /annum)

(d) De-aggregation of the UHS earthquake scenarios, each having a

specified magnitude, distance, and
The probabilistic UHS obtained through severity (described by the parameter

PSHA takes account of all of the epsilon).

earthquake scenarios that could affect
the site, and describes the ground motion The de-aggregation of the hazard varies
level that corresponds to a specified with the return period and the ground

return period or annual probability of motion period of interest, as shown in

exceedance. The UHS represents the Figure C2.14. For short return periods,
contributions from many different the hazard tends to be dominated by
earthquakes, perhaps including small smaller, more distant earthquakes (top
nearby earthquakes and more distant row of Figure C2.14) while for long
larger earthquakes, and may not be a return periods the hazard tends to be
realistic representation of the response dominated by larger, more nearby
spectrum of any individual scenario earthquakes (bottom row). For short
earthquake. For this reason, if it is ground motion periods, the hazard tends
desired to use ground motion time- to be dominated by smaller earthquakes
histories to represent the response (left side of Figure C2.14) while for long
spectrum, it is then necessary to identify ground motion periods the hazard tends
one or more scenario earthquakes that to be dominated by larger earthquakes
dominate the hazard for that return (right side). Selection of time-histories
period in the ground motion period therefore needs to take account of the
range of importance for the structure. return period of the UHS and the ground
This process, termed de-aggregation of motion period of the structure that is to
the UHS, results in one or more be analysed.

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis


Figure C2.14. De-aggregation of the hazard for peak acceleration (left) and 1 s response
spectral acceleration (right) for return periods of 500 years (top) and 10,000 years
The de-aggregation of the UHS is used spectral matching of time-histories to

whenever it is necessary to represent the represent the design ground motions.

design response spectrum by a single
earthquake. This is done whenever (e) Scenario Spectrum in Place of the UHS
dynamic analyses of a structure are
As discussed above the UHS represents
performed using a ground motion time-
the contributions from many different
earthquakes, perhaps including small
We next consider the suitability of the nearby earthquakes and more distant
UHS as a representation of the response larger earthquakes, and may not be a
spectrum of a single earthquake time- realistic representation of the response
history. We show that there are two spectrum of any individual scenario
sources of over conservatism in the use earthquake. For example, small nearby
of the UHS as a representation of a earthquakes are expected to have
single earthquake time-history. In both relatively high ground motion levels at
cases, the UHS is too “broadband” (it is short periods and relatively low ground
large over a range of periods that is motion levels at long periods.
unrealistically broad), rendering it sub- Conversely, large distant earthquakes are
optimal for use as a target for scaling or expected to have relatively low ground
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis
motion levels at short periods and implies that large-amplitude spectral
relatively high ground motion levels at values will occur at all periods within a
long periods. single ground motion time-history. An
alternative, termed a Conditional Mean
Instead, it is preferable to use ground Spectrum (CMS), provides the expected
motion time-histories that are scaled or (i.e., mean) response spectrum,
spectrally matched to a set of scenario conditioned on the occurrence of a target
spectra representing a small set of spectral acceleration value at the period
magnitude and distance combinations of interest (Baker and Cornell, 2006;
that dominate de-aggregation of the Baker, 2011). Baker (2011) shows this
UHS. These scenario spectra are in turn spectrum to be the appropriate target
scaled so that they are enveloped by the response spectrum for the goal described
UHS and collectively represent that above, and it is thus a more appropriate
broadband spectrum. One scenario from target for scaling and spectrally matching
a small nearby earthquake may represent ground motion time-histories as input to
the short period part of the UHS, while dynamic analyses. Baker (2011)
another scenario from a larger more demonstrates that the CMS spectrum
distant earthquake may represent the

maintains the probabilistic rigour of
long period part of the UHS. PSHA, so that consistency is achieved
between the PSHA and the ground
(f) Conditional Mean Spectrum in Place of
motion selection. This enables
the UHS
quantitative statements to be made about
There is another source of over the probability of observing the structural
conservatism in the UHS that may response levels obtained from dynamic
become important when the scaled analyses that use this spectrum; in
scenario spectrum lies considerably contrast, the UHS does not allow for
above the median level for that such statements (Baker, 2011). Figure
earthquake scenario. Baker (2011) C2.15 shows an example.

showed that the UHS conservatively


ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis

Figure C2.15 Left: Median (solid line) and median plus two standard deviations
(dashed line) design spectrum for a magnitude 7 earthquake at a closest distance of 12

km, shown with recorded response spectra (green), one of which (blue) is close to the
design spectrum at a period of 1 second. Right: Derivation of the conditional mean
spectrum for a period of 1 second from a scenario spectrum. The conditional mean is
derived from the correlation between response spectral values at adjacent periods.
(Baker, 2010).
(g) Vertical Response Spectrum. from the PSHA, and then scaling that
spectrum by the vertical/horizontal ratio
In dam engineering, the horizontal of a model such as Gulerce and
components of ground motions usually Abrahamson (2011).
dominate the design. Accordingly, the

vertical component of ground motion that C2.5 Requirements of a Seismic Hazard

is used in design should be compatible Assessment
with the horizontal components; it should
represent the characteristics of the
vertical component that are expected to 1. Distributed Earthquake source models.

occur in conjunction with the horizontal In current earthquake source models for
components. Therefore, the vertical
Australia, distributed seismicity is
component is not derived from the PSHA modelled in quite different ways that give
– that would represent earthquake
rise to significant epistemic uncertainty.
magnitudes and distances that are Distributed seismicity is represented by
generally quite different from those of the
discrete seismic source zones (Brown and
horizontal component. Instead, the Gibson, 2004); by spatially smoothed
vertical component is derived from the
seismicity (Somerville et al., 2009), and
de-aggregation of the horizontal by a layered seismicity approach
component so that the vertical response
(Burbidge and Leonard, 2011), as
spectrum represents the same described in more detail below. It is
combination of earthquake magnitude
necessary to include these alternative
and distance as does the dominant source models in the seismic hazard
contribution to the horizontal spectrum.
analysis in order to account for this
This is accomplished by calculating the epistemic uncertainty. None of these
horizontal response spectrum using the
three source models contains any fault
de-aggregated magnitude and distance sources; these need to be treated
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March
2017 76
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis
separately. ground zones, a set of regional zones
having higher seismic activity rates than
ES&S AUS5 Seismic Source Zone Model the background zones, and hotspots that
contain concentrations of earthquake
The Environmental Systems and Services
activity within regional source zones. The
(ES&S) AUS5 source zone model
2012 Australian Seismic Hazard Map
originally developed by Brown and
(Burbidge and Leonard, 2011) is
Gibson (2004) uses geological and
composed of three layers of seismicity.
geophysical criteria in combination with
historical seismicity to identify zones of • Background Source Zones. There are
uniform seismic potential, and then uses two of these, one for the cratonic
historical seismicity to characterise the region of the western part of
seismic potential of each zone by means Australia, and another for the non-
of the a-values and b-values of the cratonic region of the eastern part of
Gutenberg-Richter earthquake recurrence Australia. For each region there are
model, together with an estimate of the two models: one that reflects the
maximum magnitude of earthquakes in actual data and another that uses a
each zone. This approach has the minimum value (floor).

advantage of allowing for the
incorporation of geological and • Regional Source Zones. These
geophysical information as well as represent the long-term earthquake
activity above the background level in
seismicity data in the identification of
seismic source zones. specified broad regions
• Hotspot Source Zones. These
Risk Frontiers Spatially Smoothed represent the short-term earthquake
Seismicity Model activity, including earthquake swarms
Judgment is required in defining the and clusters, above the background
source zone boundaries of models such as and regional source zones, that have
been occurring recently in local zones

Brown and Gibson (2004), and it is

unclear what would cause abrupt changes Burbidge and Leonard (2011) propose a
in seismicity levels across source zone number of different ways in which the
boundaries. These considerations hazard maps from each layer can be

motivate the use of spatially smoothed combined. One way is to use the highest
historical seismicity to define the hazard value from among the three
earthquake forecast, developed by Risk different layers at each site. In this
Frontiers (Hall et al., 2007). This approach, the total integrated seismic
approach gives a spatially continuous moment across Australia is not
source model without boundaries except conserved, but the hazard in the areas of
in b-value. The spatial smoothing low seismic activity is allowed to
approach has the advantages of simplicity increase without decreasing the hazard in
and of avoiding uncertainty in the areas of high seismic activity. This
geological definitions of zones, but has approach may be justifiable for
the disadvantage of not making use of application to the new building code
potentially informative geological data. provisions (AS1170.4), but it is not
suitable for site-specific studies such as
GA Layered Seismicity Model those for dams, where rigorous
Leonard et al. (2012) developed an probabilistic analysis of hazard and risk
earthquake source model for Australia is required. In applications to dams, it is
that is based entirely on historical preferable to use the version of the
seismicity. They identify a set of back Background Source Zones that does not
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis
contain a floor (minimum value), GA model has maximum magnitudes ranging
together with the Regional Source Zones. from 7.3 to 7.7.
If any of the Hotspot Source Zones
contributes significantly to the seismic As shown in Figures C2.16 and C2.17 the
hazard at the site, they should also be models can give significantly different
included as source zones. seismic hazard in some areas in Australia,
and there is no systematic outcome, with one
The three distributed earthquake source or the other model giving higher hazard. In
models all specify the maximum earthquake view of this any seismic hazard assessment
magnitudes in their recurrence models. The which uses only one source model is
ES&S and RF models both assume a potentially unreliable.
maximum earthquake magnitude of 7.5. The


Figure C2.16. Example of differences in event rates for seismic hazard assessment from AUS5
discrete seismic source zones model and Risk Frontiers spatially smoothed seismic source model.

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis

Example Dam – Equal Hazard Spectra – 5% Damping

Figure C2.17 Seismic Response Spectra using the AUS5 and GA discrete seismic source zone
models and the Risk Frontiers spatially smoothed seismic source model, and their average.
2. Fault sources. varies in completeness because sampling
is biased by the available data bases, the
(a) Active and Neotectonic Faulting in extent of unconsolidated sedimentary
Australia cover, and the relative rates of landscape
and tectonic processes.
Fault sources here means active faults
and neotectonic faults as defined in Following Clark et al. (2011) and

Section 2.2. Sandiford et al. (2004), Seismologists

consider neotectonic faults as potentially
The long term uplift and subsidence active if they have undergone
caused by active faulting in geological displacement under the current stress

time has a significant effect on drainage regime in Australia, and hence may have
patterns in many parts of Australia, with the potential for displacement in the
the result that dams are frequently located future. The age of the current stress
in close proximity to active or regime in Australia is estimated to lie in
neotectonic faults. the range of 5 to 10 million years
(Sandiford et al., 2004).
An Australia-wide assessment of faulting
was made by Clark et al. (2011; 2012). In Australia, geological maps typically
They analysed a catalogue of over 200 show numerous faults but do not indicate
neotectonic features, 47 of which are whether they are active in the current
associated with named fault scarps. The stress regime; most of them are probably
data were derived from analysis of not. For example, if these faults were
DEMs, aerial photos, satellite imagery, previously active under a different stress
geological maps and consultation with orientation, it is possible that they are
state survey geologists and a range of unfavorably oriented to undergo slip
other geoscientists. Verifying the features under the current stress orientation.
as active as defined for neotectonic faults However, if they are favourably oriented,
is an ongoing process. The catalogue then consideration should be given to the
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis
possibility that they have been reactivated event within the past 35,000 years or may
under the current stress regime. have the potential to generate a large
earthquake within a 35,000 year time
Clark et al (2011) list a number of faults frame, and so may fit the definition of
which have been investigated and found active fault that these guidelines has
to have had multiple surface rupture adopted. Figure C2.18 summarizes data
events over intervals of hundreds of for some of these faults which have
thousands of years, with intervals of a resulted in surface rupture. Others which
few tens of thousands of year between did not result in surface rupture are listed
event. In some cases these faults undergo within the text of Clark et al (2012).
a long period of quiescence, and may
currently be in a quiescent phase.
However, others may have had a large


Figure C2.18 Compilation of surface breaking earthquake recurrence data. (Clark et al. 2012)

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis
(b) Alternative Approaches to seismicity is not represented in the
Incorporating Active and Neotectonic distributed source zones, consistent with
Faults the Characteristic earthquake recurrence
model, which is described next. This
In view of the short time span of the approach requires an estimate of the slip
historical earthquake catalogue in rate of the fault. The episodic nature of
Australia, it is unclear how best to large earthquake ocurrence, described
incorporate active and neotectonic faults below, necessitates the careful
into the earthquake forecast. consideration of the time period over
which the slip rate of the fault should be
Brown and Gibson (2000) proposed a
method which subtracts fault-related
seismicity from the area source zone in (c) Alternative Earthquake Recurrence
which the fault occurs, and insert a fault models
source having that seismicity, using a
Gutenberg-Richter recurrence model. The distribution of earthquake
This approach assumes that the fault magnitudes in distributed earthquake
seismicity is represented in the source zones is usually assumed to follow

background seismicity of the area source. the Gutenberg-Richter model (Figure
C2.19, top left). However, the
An alternative approach (Somerville et al, distribution of earthquake magnitudes on
2008) is to add a fault source whose
discrete active and neotectonic faults may
seismicity is based on slip rate, without be better represented by the characteristic
modifying the background seismicity of recurrence model (Schwartz and
the area source, using a characteristic Coppersmith, 1984), in which most of the
earthquake recurrence model. This fault slip is taken up in large earthquakes
approach assumes that the fault (Figure C2.19, bottom left and right).

Figure C2.19. Left: Interval recurrence for the Gutenberg-Richter (top) and
Characteristic earthquake recurrence models (bottom). (Wesnousky et al., 1983). Right:
Cumulative recurrence for the Characteristic earthquake recurrence model. Source:
Schwartz and Coppersmith, 1984.

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis
If the characteristic recurrence model characterised by different earthquake
applies, then the recurrence rate of large recurrence behaviour.
earthquakes may be underestimated by
the Gutenberg-Richter model that is In practice unless there is definitive
derived from historical seismicity if it information to the contrary, faults identified
only contains small earthquakes, as as active as defined in Section 2.2 should be
indicated on the right side of Figure considered still active. If there are differences
C2.19. For active faults, it is preferable of view from Seismologists as to whether a
to estimate the recurrence rate of large fault is still active this can be managed by
earthquakes from geological data, such as assigning a conditional probability to the
fault slip rates, rather than historical occurrence of the earthquake motion to
seismicity, because it is unclear whether reflect these opinions.
the earthquake recurrence of fault sources
(e) Modelling the Maximum Credible
follows the Gutenberg-Richter model.
Ground Motion at the Dam site from an
(d) Alternative Recurrence Behaviour of Active Fault
Active Faults
Where an active fault has been identified

Clark and Van Dissen (2006) and Clark which could result in significant ground
et al. (2012) found that earthquake motions at the dam site its contribution to
activity on faults in Australia is episodic, ground motion will have been considered in
the PSHA as described above.
with clusters of earthquakes on a given
fault occurring close together in time
As detailed in Table 2.1 of Section 2.6, for
(several tens of thousands of years),
deterministic seismic hazard for Extreme
separated by longer periods (several
Consequence Category dams the maximum
hundreds of thousands of years) of no
credible ground motions which might occur
large earthquake activity. This is
at the dam site from rupture of active faults
inconsistent with the random temporal
are required
(Poisson) distribution of earthquakes that

is usually assumed in seismic hazard There are however uncertainties in doing this
analysis. Using the results of Clark et al. including whether the whole fault or only part
(2006), it may be possible to identify ruptures: the magnitudes of the resulting
which faults are currently in an active earthquakes which might occur; the location

phase and which are currently in an of the focus of the earthquake; the ground
inactive phase. This could then be applied motion models to be adopted. These should
to the evaluation of the seismic potential be discussed with the Seismologist so the
of active faults in seismic hazard Owner and Consultant understand the
evaluations. potential degree of conservatism inherent in
the ground motion assessment.
Clark et al. (2011) reviewed knowledge
pertaining to the seismogenic deformation of It should be noted that in cases where the
the Australian continent over the last 5-10 Ma active fault is close to the dam site these
(the Neotectonic Era). Based upon perceived ground motion estimates are likely to be
differences in character of the seismogenic greater than obtained from PSHA. This is
faults across the continent, and guided by because the PSHA will have assigned an
variations in the geologic and geophysical annual probability to the occurrence of
makeup of the crust, they propose six onshore earthquakes on the fault while the process
neotectonic domains. A seventh offshore above takes no account of this. The process
domain was defined based upon analogy with can be considered as equivalent to Probable
the eastern United States. These domains are Maximum Flood estimate which also has no
annual probability assigned to it.
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis
3. Faults in the dam foundation develop ground motion prediction models
for Australia. In view of the significant
In practice only active faults are likely to differences in the ground motions that are
be of significance to the dam. However predicted by the different ground models,
there may be neotectonic faults which it is necessary to include alternative
warrant further investigation to determine ground motion prediction models in the
if they are active. seismic hazard analysis in order to
account for epistemic uncertainty in
4. Ground motion prediction models.
which of these models is more applicable
There are insufficient strong motion in Australia. The following briefly
recordings from earthquakes in stable summarizes the available models (in
continental regions anywhere in the 2016).
world, including Australia, to form a
Liang et al. (2008).
basis for the development of ground
motion prediction models using Liang et al. (2008) estimated strong
regression analysis of recorded strong ground motions in southwest Western
ground motions. Such analyses have Australia using a combined Green's

only been feasible for crustal earthquakes function and stochastic approach. This
in tectonically active regions, and while model is applicable to the Yilgarn craton,
the ground motion prediction equations and may also be applicable to other
derived this way have been used
cratonic regions of Australia, but is not
extensively in Australia, the applicability applicable to non-cratonic regions,
of these models in Australia is still not including Perth. This model uses
well established. Accordingly, recent epicentral distance as the distance
investigators (Liang et al., 2008; measure. Figure C2.20 shows the crustal
Somerville et al., 2009; and Allen, 2011), domains in Australia as defined by Clark
described further below, have used et al (2012).
seismologically based methods to

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis

Figure C2.20 Neotectonic domains map of the Australian continent. Neotectonic features
represent known or suspected features, primarily topographic fault scarps. (Clark et al,
Somerville et al. (2009). Cratonic Australia and Non-Cratonic
Australia. The cratonic regions of
Somerville et al. (2009) demonstrated Australia include much of Western
their ability to simulate the recorded

Australia (but not the coastal strip west of

ground motions of small earthquakes that the Darling Fault, including Perth); south-
occurred in Eastern and Western central South Australia (including the
Australia, and developed earthquake site); the northern part of the Northern
source scaling models for Australian Territory; and north western Queensland

earthquakes based on earthquake source (Clark et al, 2011). Non-Cratonic

modelling of the Mw 6.8 1968 Meckering Australia consists of the remainder of
and the Mw 6.25, 6.4 and 6.6 1968 Australia, including Eastern Australia and
Tennant Creek earthquakes. They then part of the coastal margin of Western
used a broadband strong ground motion Australia, which includes all of the state
simulation procedure based on the elasto- capital cities, is on Non-Cratonic
dynamic representation theorem and Australia.
Green’s functions calculated from crustal
structure models for various regions of Allen (2012).
Australia to calculate ground motions for
earthquakes in the magnitude range of 5.0 Allen (2012) developed a ground motion
to 7.5. These ground motions were then model for south eastern Australia based
used to develop ground motion prediction on the stochastic model, having calibrated
equations, which were checked for the parameters of the stochastic model
consistency with available data from using recordings of small earthquakes in
Australian earthquakes at each step. These south eastern Australia (SEA). The source
ground motion models predict response and attenuation parameters provided in
spectra for two crustal domain categories: Allen et al. (2007) were first reviewed and
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March
2017 84
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis Method
modified in light of additional available (Abrahamson and Shedlock, 1997).
data and more rigorous statistical analysis.
The dependence of stress drop on Selection of Ground Motion Models for
earthquake depth was examined, and use in Australia
options were provided for variable stress
The NGA West 2 models are based
parameter values in the ground motion
mostly on a large global set of recorded
prediction equation. The near-surface,
ground motion data from tectonically
path-independent diminution parameter
active regions, but none of those data are
immediately beneath the station, k0,
from Australia. The Allen (2012) and
(Anderson and Hough, 1984; Campbell,
Somerville et al. (2009) ground motion
2009; Van Houtte et al., 2011) was also
prediction models were both developed
examined for average station conditions
for Australia, but are not based on a large
in SEA, in addition to the parameter’s
set of recorded strong motion data from
correspondence with a limited dataset of
average shear-wave velocity
measurements in the upper 30 m (Vs30) In view of this the NGA West 2 model
at seismic recording stations across should be included.
Australia. These updated source and
attenuation parameters were used as Judgments need to be made as to the

inputs to the stochastic finite-fault relative weights that should be given to
software package, EXSIM (Motazedian the Australian and global models. This is
and Atkinson, 2005; Atkinson and Boore, a matter for the Seismologist to
2006). Five percent damped response determine, in consultation with the Owner
spectral accelerations were simulated for and the Consultant, taking account of
earthquakes of moment magnitude Mw whether the site is in cratonic or non-
4.0 to 7.5. These stochastic data were then cratonic conditions.
regressed to obtain model coefficients and
the resulting ground motion prediction It can be expected that new models will
model was evaluated against recorded be developed for Australian and relevant
International conditions and given they

response spectral data for moderate-

magnitude earthquakes recorded in south are rigorously peer reviewed, they may be
eastern Australia. included in the PSHA with appropriate
PEER-NGA West 2.

Impact of Site Conditions on Ground

These are the most recently developed Motion Level
ground motion models for shallow crustal
earthquakes in tectonically active regions. Ground motion prediction models used in
They were developed by five groups: earthquake engineering are based on three
Abrahamson et al. (2014), Boore et al. main parameters: the magnitude of the
(2014); Campbell and Bozorgnia (2014), earthquake, the distance of the earthquake
Chiou and Youngs (2014), and Idriss from the site, and the site characteristics.
(2014), and are compared by Gregor et al. It has long been known that site
(2014). In view of the great care that was characteristics have a strong influence on
put into documenting the NGA West 2 ground motion level. Until recently, site
metadata that describe the strong motion characteristics have been represented by
recordings, the vastly larger size of the broad geological categories such as
data set that has been used, and the “rock” or “soil.” In eastern Australia, it
diligence that has been applied by the has been common to assume that the site
modellers, the NGA Program has resulted characteristics of dam abutments can be
in a set of ground motion models that represented by the “rock” site category in
have a much more substantial basis than ground motion models such as Sadigh et
the earlier 1997 generation of models al. (1997). However, this ground motion

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 85
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis Method
model was found in the course of the refraction surveys which were often
NGA Project to be representative of soft carried out as part of the original site
rock sites in California having an average investigation for the dam site.
shear wave velocity (Vs30) of only 520
m/sec, while many dams in Australia may The ground motion models for Australia
be founded on hard rock having Vs30 of that were developed by Somerville et al.
1,000 m/sec or more. In view of this, re- (2009) and Allen (2011) do not have
evaluation of the seismic hazard which Vs30 as a variable, and instead are for
was carried out using Sadigh et al (1997) rock site conditions. The Vs30 used by
may identify a significant level of Somerville et al. (2009) is 865 m/sec,
conservatism in those hazard assessments which is consistent with average rock site
for those dams located on hard rock conditions in Australia. Similarly, the
foundations. Vs30 of the Allen (2011) model is
assumed to represent a Vs30 of 820
Ground motion models, such as the NGA m/sec. The site amplification model of
West 2 models (Gregor et al., 2014) have Campbell and Bozorgnia (2008) can and
been developed that quantify site should be used to adjust the shear wave
characteristics in a much more rigorous velocities represented in these models to
way. Specifically, these models specify that of a specific site where the Vs30 of

the site characteristics using Vs30, which the site is significantly different to that in
is the average shear wave velocity in the the models.
uppermost 30 meters below the ground
surface. Amplification of ground motion
is inversely proportional to Vs30. The
amplification is roughly equal to the
square root of the ratio of subsurface to
surface shear wave velocity, and
illustrated in Figure C2.21. Although
Vs30 is not yet routinely measured in the
foundation investigations for dams, it can

usually be inferred from the P-wave

velocities obtained from seismic

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 86
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis Method

Figure C2.21 Dependence of response spectral acceleration on Vs30 for a magnitude 7
earthquake at a distance of 30 km. (Abrahamson and Silva, 2008).
5. Shear wave velocity of the dam what might be expected can be obtained
foundation bedrock from Table C2.4 and / or by consulting an
experienced engineering Geologist.
The shear wave velocity can be measured
However in most cases if there are no
directly or may be assessed indirectly
data, or what data there are only old “P”
from seismic refraction “P” wave
wave data, site specific “S” wave testing
velocities. Many dams have had seismic

should be carried out because the seismic

refraction (“P” wave) surveys carried out
hazard is significantly dependent on Vs30
as part of the site investigations for the
and as can be seen from Table C2.4
dam and these may be used to estimate
unless measurements are made at the dam
site the potential range of Vs30 is large.

If there are no data for the dam, a guide to

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 87
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis Method
Table C2.4 Shear wave velocity Vs30 (m/s) site classification applicable to Australian regolith
conditions (Mcpherson and Hall, 2007)

Site Class Vs30 (m/s) Geological Materials

B >760 Fresh to moderately weathered hard rock units (Plutonic &
metamorphic rocks, most volcanic rocks, coarse-grained
sedimentary rocks Cretaceous & older)
BC 555 - 1000 Highly weathered hard rock; some Tertiary volcanics
C 360 - 760 Sedimentary rocks of Oligocene to Cretaceous age; coarse-
grained; sedimentary rocks of younger age; extremely
weathered hard rock units
CD 270 - 555 Sedimentary rocks Miocene and younger age, unless
formation is notably coarse grained; Plio-Pleistocene alluvial
units; older (Pleistocene) alluvium, some areas of coarse
younger alluvium
D 180 - 360 Younger (Holocene to Late Pleistocene) alluvium

DE 90 - 270 Fine grained alluvial, deltaic, lacustrine and estuarine
E <180 Intertidal and back-barrier swamp deposits
6. Effects of soil overlying bedrock founded on soft or deep valley alluvium, a
Current ground motion prediction models site-specific analysis may be required.
estimate the amplification of ground motions
by shallow geology using Vs30, the time- Generally speaking, the larger the value of
averaged shear wave velocity over the upper Vs30, the more likely it is that it provides an
30 metres of the ground. Vs30 is a continuous adequate representation of the amplification

variable that spans the range from soft soils, of the site while avoiding the uncertainties
stiff soils, weathered rock, soft rock, to hard associated with nonlinear soil response
rock. Vs30 is specified in the range of 180 analysis. Also, generally speaking, the lower
m/sec to 1100 m/sec (ranging from soft soil to the level of the ground motion that is input

hard rock). The data are insufficient to into the soil, the less nonlinear will be the
constrain amplification in rock above 1100 response of the soil, and the more reliable will
m/sec; it is expected to decrease slowly. be the estimate of the soil amplification using
Below 180 m/sec, the use of Vs30 is not Vs30.
viable and a site-specific soil response study However if the dam is founded on sediments
is necessary. Such soils are described in
or deep residual soils and completely
United States codes as NEHRP sites E and F. weathered rock overlying the bedrock which
Use of Vs30 to estimate soil amplification at a are expected to have significant nonlinear
specific site assumes that the average shear response, then the response of these materials
wave velocity, shear modulus reduction, and should be analysed in a separate study of
damping increase profiles represented in the nonlinear soil response using SHAKE,
strong motion data base used to develop the QUAD4M or similar programs that would not
ground motion prediction models are a be part of the seismic hazard assessment. The
reasonably accurate representation of the Consultant who performs this analysis should
specific profiles at the site. If this is not the specify the subsurface level (depth in the
case, then a site-specific analysis of the soil profile) and the associated (subsurface) Vs30,
amplification should be done as described Z1.0 and Z2.5 at which the input ground
below. For most dam sites where the dam is motions are to be provided. In this case, Vs30

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 88
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis Method
is the time-averaged shear wave velocity over ground motions. The minimum
a depth of 30 meters (m/s) below the selected magnitude considered by Burbidge and
subsurface level, and the Z1.0 and Z2.5 Leonard (2011) in the new draft seismic
(depths to shear wave velocities of 1.0 and 2.5 hazard maps of Australia for general
km/sec) are also adjusted accordingly. The building code applications is magnitude
ground motions developed to represent free- 4.5. For dams, it will be appropriate to
field ground motions at this subsurface level use a higher minimum magnitude.
are then used as input into the nonlinear
analysis. 9. Quantification and reporting epistemic
The analyses should use the non-liquefied
properties of the soil. Uncertainty in the inputs and models should
be modelled and reported. There exists
One dimensional analyses such as SHAKE epistemic uncertainty (about the true state of
are potentially conservative for 2D and 3D nature) in earthquake source models (and the
structures such as embankment dams, so historical earthquake catalogues on which
where the results are important to dam safety they are based); about how to incorporate
decision making it may be necessary to carry active faults and in the source parameters of
out 2D modelling such as QUAD4M or these faults; in ground motion prediction

QUAKEW. models; and in the shear wave velocity
assigned to the dam foundation and the
7. Response spectra damping coefficients adopted. These
uncertainties lead to uncertainty in the true
No commentary. value of the hazard at a specified return
period. These uncertainties can be significant
8. Minimum Magnitude Earthquake
(Somerville and Thio, 2011). Figure C2.22
Even small earthquakes can produce shows the results of a seismic hazard
relatively large peak accelerations at close assessment where the mean (best estimate) of
distances, but such earthquakes generally the hazard would result in the foundation
have low damage potential, and it is largely being non-liquefiable, but the 85%

desirable to exclude earthquakes below a fractile and 95th fractile would likely result in
specified minimum magnitude to avoid liquefaction, giving a totally different
such events contributing to an outcome for only marginally less likely loads.
unrealistically high hazard level,

especially for PGA and short period

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 89
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis Method
Example Dam: Equal Hazard Spectra – 10,000 yr – 5%

Figure C2.22 Example of horizontal spectral response acceleration showing the epistemic
uncertainty in the true value of the hazard. The mean hazard is shown by the red line and the 5%

and 95% fractiles are shown by the blue lines.

10. De-aggregation plots and selection of may be appropriate with a less detailed PSHA

time-history motions. as described below for significant

Earthquake de-aggregation plots showing consequence category dams used initially, and
magnitude – distance contributions should be the more detailed assessment carried out if it
provided because these are a significant input is recognised that the seismic hazard is
into analyses for assessing embankment dam critical.
deformations and liquefaction. The de-
aggregation of the hazard varies with the For significant and low consequence category
return period and with the period of the dams situations where a site specific and
ground motion, so the magnitude and distance detailed PSHA may be required include dams
should be selected based on knowledge of the which have potentially liquefiable
structure being analysed and the return period foundations, or other features such as walls
being considered. retaining the embankment which are sensitive
to the seismic loading.
11. Reservoir-induced seismicity.
C2.6 Selection of Design Seismic Ground
See Sections C2.1.5 and C2.11. Motion - Deterministic Analysis Approach
Operating Basis Earthquake
In some situations for high and extreme
consequence category dams a staged approach

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 90
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis Method
Other stakeholders may for example include a understand the characteristics of the active
local government council or other water fault. This may result in additional
supply authority which obtains their water investigations being required such as
from the dam Owner. trenching across the fault and dating the
episodes of displacement. Once the MCE
Safety Evaluation Earthquake loading is finalised either adopt it and / or
adopt a risk based approach.
In assessing the required factor of safety to
achieve a low likelihood of failure the effect If there is still some doubt about whether a
of the 85th fractile (and possibly 95th fractile fault is active or not, consult the Seismologist
for Extreme Consequence Category dams) (s) and Geologist involved and assign a
ground motions on the dam should be probability that the fault is active and apply
assessed. These are about 3 times and 10 this in the event tree logic in a risk analysis.
times less likely than the median or 50 th
fractile ground motion. Alternatively the 1 in The use of a deterministic analysis approach
30,000 AEP or 1 in 100,000 AEP ground requires that a factor of safety should be
motions may be used. applied to estimated stresses and deformations
within the dam and its foundations to give a
If the dam would fail leading to breach for low likelihood of the dam failing given the

these ground motions the likelihood of failure SEE loading. Table C2.5 shows the
would probably be too high to satisfy conditional probabilities of failure required to
ANCOLD (2003, 2016) tolerable risk criteria. just achieve ANCOLD (2003) tolerable risk
Some additional information to assist in this guidelines. To allow for the fact the tolerable
assessment is given below. risk criteria are for the sum of all potential
failure modes, and the seismic loading
For High B, High C , Significant and Low
component may only be a small part of the
Consequence Category dams, if the structure
overall probability of failure, and to give
is susceptible to liquefaction or has
some margin of safety that the risks are below
components which will fail at loads only a
the limit of tolerability, the conditional
little greater than the loads in Table 2.1, check
probability of failure for seismic loading for
the design for the critical load and assess the

existing dams, given the SEE should be about

adequacy of the design using risk assessment
1 in 100, and for new dams or major
augmentation, about 1 in 100 to 1 in 1000.
It should be noted that in Australia the MCE This requires a significant factor of safety on

loading on known active faults may be stresses or deformations for the SEE load. It
significantly greater than the probabilistic should be noted that individual risk is the
loadings in Table 2.1 even though the controlling factor for High C, Significant and
probabilistic loadings include the effects of Low Consequence Category dams.
the active faults. This is because the MCE
The requirement for a factor of safety on
approach effectively ignores the AEP of the
stresses or deformations applies to all
earthquake on the fault and this may be very
consequence categories and is a critical
low, so it does not have as much effect in the
requirement if a deterministic analysis
probabilistic assessment as it does in the MCE
approach is to be followed.
approach. In seismically active areas the
As there are no unique relationships between
reverse may apply.
factor of safety and likelihood of failure the
If the MCE loading is significantly larger than Consultant will have to apply engineering
the probabilistic loading detailed discussions judgement to achieve the required low
should be held with the Seismologist to better likelihood of failure.

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 91
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis Method
Table C2.5 Indicative required conditional probabilities of failure given the ANCOLD (2013)
Design Earthquake
Dam ANCOLD ANCOLD ANCOLD Conditional Conditional
Consequence (2013) (2012) (2003) Probability of Probability of
Category, or Design of Consequence Societal Failure Failure
“class” Concrete Categories Risk To achieve To achieve
Gravity Dams Potential Annual ANCOLD(2003) ANCOLD(2003)
Recommended Life Loss Probability Limit Criteria Limit Criteria
Design Loads (PLL) of Failure (a) (b) (c)
Limit (a)
Extreme 1:10,000 AEP >50 2E-05 <0.2 <0.1
100 1E-05 <0.1 0.05
1000 1E-06(d) <0.01 0.005
High A 1:10,000 AEP 1 to 50 1.0E-03 to < 0.2 < 0.1
High B 1:5,000 AEP 0.1 to 5 1.0E-02 to <1 <0.5
2.0E-04 (0.5,0.05) (< 0.25 to

High C 1:2,000 AEP <0.1 to 5 1.0E-02 to <0.4 <0.2
(<0.2,0.02) (<0.1, 0.01)
Significant 1:1,000 AEP <0.1 to 1 1.0E-02 to <1 (<0.5)
(<0.1, (<0.05, 0.0005)

Low 1:1,000 AEP <0.1 1.0E-02 <1 (<0.1)

(<0.1, (<0.005, 0.0005)

(a) The figures include the contribution to the annual probability of failure from other failure modes.
(b) Figures in brackets controlled by individual risk criteria, 1.0E-04 / annum for existing dams;

1.0E-05 for new dams or major augmentations. Figures assume vulnerability of the person most at
risk = 1.
(c) These figures assume that the contribution to the risk from seismic loads is half the total.
(d) This assumes that the horizontal truncation in the societal risk plot is removed in ANCOLD

C2.7 Selection of Design Seismic Loading- walls; e.g. between spillway and
Risk Based Analysis Approach embankment; failing.
For the risk analysis the seismic loads should (c) The AEP of the earthquake loading
be partitioned taking account of the fragility which is likely to result in
of the structure. For example: displacement of a concrete gravity
(a) If liquefaction is an issue, the
earthquake loading with an AEP at The uncertainty in the earthquake loading
which liquefaction is widespread should be considered and modelled in the risk
would be used as a partition boundary. analysis.
(b) The AEP of the earthquake loading
which results in critical retaining

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 92
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis Method
C2.8 Modelling vertical ground motions At least four or five ground motion records
Gulerce and Abrahamson (2011) provide are generally needed for advanced
ground-motion prediction equations (GMPEs) deformation analyses. A relatively large suite
for the vertical-to-horizontal spectral of records is required due to the range of
acceleration (V/H) ratio, and the methods for deformation predictions that may occur even
constructing vertical design spectra that are for a carefully selected set of motions. Since
consistent with the probabilistic seismic only 4 or 5 records are being used, the intent
hazard assessment results for the horizontal of the study is not to define the full range of
ground motion component. potential displacements but to determine the
average, expected response for the specified
The proposed V/H ratio GMPE is dependent level of earthquake loading.
on the earthquake magnitude and distance,
and accounts for the differences in the non- Three ground motion components should be
linear site-response effects on the horizontal provided for each record: two horizontal and
and vertical components. This results in large one vertical. While vertical motions are often
V/H ratios at short spectral periods for soil considered to have a modest effect on
sites located close to large earthquakes. deformation predictions, advances in
developing appropriate and consistent
It is suggested that this method be used rather motions and the ease with which they can be

than using a constant ratio of vertical to included in many sophisticated analyses
horizontal ground motion as has been warrants their routine use in deformation
commonly done. analyses.
C2.9 Selection of Response Spectra and The suite of records should be obtained from
Time-History Accelerograms different source earthquakes to reduce
The amount of damping assumed for the unintended bias in the record selection. The
response spectrum and analyses will depend following criteria may be considered in the
on the type of structure. For embankment selection of earthquake record:
dams the amount of damping will depend on a) Records to be used for preparation of
the extent of straining within the dam. For

site specific time-histories should

example, the Makdisi-Seed method for originate from a seismic event similar
estimating earthquake induced deformations
to the target design earthquake (e.g.,
in an embankment varies both the bulk magnitude, fault distance, and focal
modulus and the damping factor according to

depth). The site condition for each

the shear strain in the embankment. For record should reasonably correspond
concrete dams, the USACE (1999) states:
to the site condition for the target
“Energy dissipation in the form of a damping
response spectrum. For example, it
ratio is included as part of the response
may be appropriate to use a record
spectrum curves. For the linear elastic or
from a shallow, stiff soil site to
nearly elastic response during an OBE event,
represent soft rock conditions, but not
the damping value should be limited to 5
a deep soil record.
percent. For the SEE excitation, a damping
b) The shape of the response spectrum
constant of 7 or 10 percent may be used
for each record should reasonably
depending on the level of strains and the
match the target response spectrum
amount of inelastic response developed in the
over the frequency range of interest.
This frequency range may be rather
Suitable methods can be used to convert a large and will typically include low
response spectrum to account for higher frequencies (long periods).
damping factors than the one for which the c) Scaling factors may be applied to the
response spectrum was prepared (e.g. Rezaein record to provide a best fit to the
et al., 2014a, b). response spectrum over the period
range of interest (see Figure C2.23).
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 93
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis Method
Alternatively, spectral matching motions, such as Arias Intensity or
programs such as RSPMatch can be significant duration. Attenuation
used to more closely match the target relationships are available for these
response spectrum over a wide range parameters allowing their inclusion in
of frequencies. deterministic or probabilistic hazard
d) Spectral matching techniques should estimates (e.g., Watson-Lamprey and
be carefully applied to preserve as Abrahamson, 2006; Travasarou, et al.,
much of the original character of the 2003; Kempton and Stewart, 2006].
earthquake record as possible (e.g., f) Original earthquake records can be
relative magnitude and duration of obtained from a number of online
velocity peaks). Although scaling sources, including the COSMOS and
factors are traditionally limited to PEER websites. Synthetic
values between 0.5 and 2.0, values accelerograms should be considered
outside of this range may be permitted when the design earthquake is not
in some cases (Watson-Lamprey and well-represented by the database of
Abrahamson, 2006). available records.
e) Additional criteria can be useful in
defining an appropriate suite of ground


Figure C2.23 Scaling of the input time-history for best fit with target response spectrum over
period range of interest (Perlea and Beaty, 2010).

C2.10 Earthquake Aftershocks saturate from leakage through the damaged

This is more likely to be an issue where an face slab or membrane after the main
active fault or individual neotectonic fault is a earthquake, leaving it with a low static factor
major contributor to the seismic hazard. of safety and susceptible to much larger
deformations and damage during the
An example of a dam which might be aftershock.
susceptible to aftershock seismic ground
motions is a concrete face, asphalt face or Concrete dams and appurtenant structures
other membrane face rockfill dam which has such as walls retaining the embankment
poor drainage capacity, so the rockfill may which may experience significant
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 94
Commentary – 2.0 Assessment of Design Seismic Ground Motions Analysis Method
displacement during the primary ground geology and seismicity suggest that the area
motion, sufficient to compromise the could be subject to reservoir-triggered
foundation drainage or passive or post seismicity.
tensioned anchors should be considered for
aftershocks. Depending on the dam location and prevailing
seismotectonic conditions, the RTE may
C2.11 Seismic Ground Motions from represent ground motion less than, equal to, or
Earthquakes Induced by the Reservoir greater than the OBE ground motion. RTE
The following is adapted from ICOLD ground motion should in no case be greater
(2016a). than the Safety Evaluation Earthquake ground
motion because the faults considered capable
The Reservoir-Triggered Earthquake (RTE) of triggering seismicity should be taken into
represents the maximum level of ground consideration during the seismic hazard
motion capable of being triggered at the dam evaluation. However the result might be the
site by the filling, drawdown, or the presence premature triggering of seismic events due to
of the reservoir. ICOLD Bulletin 137 on the impounding of the reservoir that would
Reservoirs and Seismicity provides the state- have occurred naturally at some longer time in
of-knowledge on reservoir-triggered the future. It is therefore justified in the case
seismicity. Section C2.1.5 gives some details. of larger dams and storages located in

seismically active regions and regions with
While there exist differences of technical
opinion regarding the conditions which cause high tectonic stresses to install a micro-
seismic network and to monitor the seismicity
reservoir-triggered seismicity, it should be
considered as a credible event if the proposed prior to, during and after impounding.
reservoir contains active or neotectonic faults This has been done on some Australian dams,
within its hydraulic regime and if the regional e.g. Thomson and Dartmouth, and micro-
and local geology and seismic record within seismicity was detected associated with faults
that area are judged to indicate potential for in the reservoir as discussed in Section
reservoir-triggered seismicity. Even if all the C2.1.5.
faults within a reservoir are considered

tectonically inactive, the possibility of For existing dams there is unlikely to be any
reservoir-triggered seismicity should not be RTE from reservoirs which have been filled
totally ruled out, if the local and regional for several decades.

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 95
Commentary – 3.0 Assessment of Embankment Dams for Seismic Ground Motion

C3 ASSESSMENT OF EMBANKMENT (c) Post-earthquake deformations for

DAMS FOR SEISMIC GROUND liquefied conditions. These methods
assume that the deformations are
primarily caused by gravitational
C3.1 Effect of Earthquakes on forces acting on an embankment
Embankment Dams. following the earthquake, and allow
No commentary. for the reduction in strength caused by
liquefaction or strain weakening of
C3.2 Defensive Design Principles for other soils. They may be carried out
Embankment Dams. using limit equilibrium and or static
The concept of ‘defensive design’ of numerical methods.
embankment dams for earthquake was (d) Advanced numerical methods for
developed by Sherard (1967) and Seed (1979) estimating deformations during and
and endorsed by Finn (1993), ICOLD (1986, post-earthquake for non-liquefied and
1999a) and ANCOLD (1998). liquefied conditions. These cover a
wide range of sophistication and
C3.3 Seismic Deformation Analysis of complexity and include dynamic
Embankment Dams. analyses using total and effective
stress methods, linear and non-linear

C3.3.1 The Methods Available and When
models, and varying degrees of
to use them
refinement of how pore pressures are
There are a number of methods available for
developed and coupled to
estimating the deformations which may occur
in embankments and their foundations during
and post seismic loading. These can be Perlea and Beaty (2010) give an overview of
summarized as: the methods and their application.
(a) Screening and empirical database The Guideline and the information presented
methods. These are applicable to below give guidance on which methods
embankments and their foundations should be used.

which do not liquefy, or experience

significant loss of strength due either It should not be assumed that the screening
to build up of pore pressure or strain and simplified methods are conservative and
weakening. These should only be where estimated deformations are near to the

applied if the criteria listed in Section tolerable deformations more advanced

C3.3.2 are satisfied, and should only methods should be used or measures taken to
be relied upon if the estimated reduce deformations.
deformations are much less than what
is tolerable; e.g. crest settlements are For risk based methods these calculations will
much less than the available freeboard. in principle need to be done for a range of
(b) Simplified methods for estimating seismic loads. In practice for many dams
deformations during earthquakes. deformations will be very small even for the
These methods are also only largest loads so it will be unnecessary to do
applicable to embankments and their the calculations for the smaller loads.
foundations which do not liquefy. For deterministic methods only the OBE and
They assume the post-earthquake SEE load will need to be analysed subject to
deformations are negligible and the the qualifications in Sections 2.6 and C2.6.
deformations during the earthquake
are due to the action of the horizontal The screening, database and simplified
inertia forces induced by the methods are not applicable to tailings dams
earthquake. They are commonly based constructed from hydraulically placed tailings
on the Newmark (1965) principle. because the tailings are not compacted to the
degree required for conventional earthfill and
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 96
Commentary – 3.0 Assessment of Embankment Dams for Seismic Ground Motion
rockfill, and will as a result be more subject to would be harmed by small movements
densification during seismic loading, and may of the embankment.
be subject to pore pressure generation and 8. There have been no historic incidents
liquefaction. at the dam that may indicate a
limitation in its ability to survive an
C3.3.2 Screening and Empirical Database earthquake.
The words in italics have been added for
Screening Method clarity.
Perlea and Beaty (2010) report that the This approach may be used as a screening
USACE do not require deformation analyses method. However for the SEE most dams in
for low to moderate height dams < 60m high Australia will have a PGA greater than 0.2g
if all the following criteria are met: (0.35g).
1. Dam and foundation materials are Empirical Database Methods
dense, not subject to liquefaction, and
do not include sensitive clays. There are a number of these methods which
2. The dam is well built and densely are developed by gathering deformation and
compacted to at least 95% of the other data on embankment dams which have

laboratory maximum dry density, or to been subject to earthquake, and relating the
a relative density greater than 80%. amount of deformation and cracking to the
3. The slopes of the dam are 3:1 (H:V) or earthquake loading experienced by the dam.
flatter, and/or (the slopes are steeper Of these the Swaisgood (1998) method is
but) the phreatic line is well below the based on a large database and is referred to in
downstream face of the embankment. Perlea and Beaty (2010) as being used by
4. The predicted peak horizontal ground USACE, and is in use in Australia.
acceleration (PGA) at the base of the
embankment is no more than 0.20g. The Pells and Fell (2002, 2003) method uses a
Compacted clay embankments on rock larger database and also records the amount of
damage as evidenced by longitudinal and

or stiff clay foundations may offer

additional resistance to deformations. transverse cracking observed. It also gives
Somewhat higher allowable PGA guidance on for what seismic loading
values may be justifiable for these (measured by Mw and PGA) transverse
dams on a case-by-case basis, cracking can be expected. This is important

although the PGA criterion should not for considering concentrated leak erosion
exceed 0.35g (USBR, 1989). following the earthquake.
5. The static factors of safety for all
Both methods are described in Fell et al
potential failure surfaces involving
(2005, 2015). They only apply to dams where
loss of crest elevation (other than
there is no potential for liquefaction or
shallow surficial slides) are greater
significant strain weakening. They should be
than 1.5 under the loading and pore-
applied conservatively, e.g. use upper bound
pressure conditions expected
estimates, allowing for the scatter in the data
immediately prior to the earthquake.
used to form the plots. They are for many
6. The freeboard at the time of the
dams quite sufficient provided the dams are
earthquake is at least 3 to 5 percent of
well constructed and there is a large margin
the dam height plus alluvial
between estimated settlement and the
foundation, and not less than 0.9 m.
Special attention should be given to
the presence and suitability of filters
for dams with modest freeboard.
7. There are no appurtenant features
related to the safety of the dam that

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 97
Commentary – 3.0 Assessment of Embankment Dams for Seismic Ground Motion
Hynes-Griffin and Franklin (1984) Pseudo- with a seismic coefficient of one half of the
Static Seismic Coefficient Method. peak ground acceleration at the base of the
dam, would assure that deformations would
This method is based on Newmark sliding not exceed 1 metre. Their suggested method
block analyses of 349 horizontal components is:
of natural earthquakes and 6 synthetic records
for a range of yield accelerations. From this (a) Carry out a conventional pseudo-static
permanent displacements u in cm were plotted stability analysis using a seismic
versus N/A where N is the yield acceleration coefficient equal to one-half the
and A is the peak value of the earthquake predicted peak ground acceleration.
acceleration at the base of the dam. (b) Use a composite S-R strength
envelope (Effective stress strength at
Hynes-Griffin and Franklin (1984) then used low stresses, undrained strength at
an arbitrary limit of 100cm (one metre) high stresses) for pervious soils and R
displacement as representing tolerable undrained strength for clays,
displacements, and from Figure C3.1, used the multiplying the strength in either case
upper bound plot of displacements, which by 0.8.
gave N/A = 0.17. After allowing for (c) Use a minimum factor of safety of 1.0.
amplification factor of 3 between the base

acceleration and that at the crest of the dam,
they concluded that a factor of safety of 1.0

Figure C3.1. Permanent displacement u versus N/A based on 354 accelerograms (Hynes-
Griffin and Franklin, 1984)
It should be noted that this method cannot be In practice for most dams with static factors
used in reverse. That is, just because the of safety about 1.5, the SEE PGA will result
factor of safety is less than 1.0, it should not in factors of safety using this method less
be assumed that the deformations will be than 1.0 so the method is not particularly
greater than 1 metre. This is because the useful. It is included here because the
upper bound deformation curve in Figure background to the method is often not
C3.1 has been used in the method. appreciated by practitioners, and to avoid its
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017 98
Commentary – 3.0 Assessment of Embankment Dams for Seismic Ground Motion
C3.3.3 Simplified Methods for Estimating are determined the permanent displacements
Deformations during Earthquakes can be calculated by double integrating the
These methods are all based on the acceleration history above the yield
Newmark (1965) approach. He introduced acceleration as shown in Figure C3.2.
the basic elements of a procedure for
evaluating the potential deformations of an Newmark’s approach is limited in
embankment under earthquake loading. In application to compacted clayey
this procedure, sliding of a soil mass along a embankments and dry or dense cohesionless
failure surface was likened to slipping of a soils that experience very little reduction in
block on an inclined plane. Newmark (1965) strength due to cyclic loading. The
envisaged that failure would initiate and approaches using this principle should not be
movements would begin when the inertia applied where embankments or their
forces exceed the yield resistance, and that foundations are susceptible to liquefaction or
movements would stop when the inertia strain weakening because they will
forces were reversed. Thus, he proposed that significantly underestimate displacements.
once the yield acceleration and the
There are a number of methods based on the
acceleration time-history of a slipping mass
Newmark (1965) principle:


Figure C3.2. Double integration method for determination of the permanent deformation period
of an embankment (Newmark, 1965).
Makdisi and Seed (1978) Method embankments subjected to earthquake
The Makdisi and Seed (1978) approach is
based on Newmark’s method, but modified The approach has been widely used and
to allow for the dynamic response of the accepted among practicing engineers.
embankment as proposed by Seed and However it is limited in application to dams
Martin (1966). The approach was developed not susceptible to liquefaction or strain
from a series of deformation analyses weakening in the embankment or its
performed on a large number of foundations. Perlea and Beaty (2010) also
warn the method may not be appropriate for
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 3.0 Assessment of Embankment Dams for Seismic Ground Motion
severe ground shaking because it does not using the methods described in
model strength losses. They also point out it Section C3.4.5.
was based on few earthquake records. (c) For potentially liquefiable soils with
a factor of safety against liquefaction
Bray and Travasarou (2007) used this greater than 1.0, determine the
method as well as their own (described residual excess pore pressure as
below) and for the dams they analysed the detailed in Section C3.4.6.1.
Makdisi and Seed (1978) method generally (d) For clay soil zones in the
underestimated the actual displacements. embankment and foundation assign
strength and pore pressure consistent
Bray and Travasarou (2007) Method.
with the soil’s behaviour in static
A simplified procedure was proposed by loading after being cracked and
Bray and Travasarou (2007) for estimating disturbed by the earthquake. If the
earthquake induced permanent clay is contractive in nature, use
displacements in earth dams using a undrained strengths. If it is dilative
Newmark-type model. This procedure is on shearing, use effective stress
based on the results from a set of simplified strengths c′, f′. Usually there will be
non-linear analysis using 688 recorded some cracking and loosening, and if
so fully softened strengths, would

ground motions from 41 earthquakes. The
flexibility of the dam system, and the apply.. Some apply an arbitrary 10%
interaction between yielding and seismic or 15% loss of strength.
loading, were considered by using a non- (e) For well compacted free draining
linear coupled stick-slip deformable sliding rockfill filters and dense sands and
model. The flexibility of the dam structure is gravels adopt the effective stress
captured through an estimate of the initial strengths c′, f′, with no change in the
fundamental period Ts. pore pressures. If large deformations
are expected in the earthquake the
Key parameters of this procedure are the dense granular materials may have
yield acceleration ky (in g), the initial loosened and will have a strength

fundamental period of the embankment, Ts, approaching the critical state

and the value of spectral acceleration for a strength, rather than the peak
damping of 5% and a degraded response strength. In practice this can be
period equal to 1.5Ts. accommodated by a small reduction

from the expected peak strengths.

Bray and Travasarou (2007) claim the
method provides improved characterisation The analysis is done with conventional limit
of the uncertainty involved in the estimate of equilibrium analysis methods. No loading
seismic displacement. It can be also be from the earthquake is applied, since this is a
incorporated into a probabilistic framework. post-earthquake analysis.
C3.3.4 Post-earthquake Deformations for If the liquefied zone is subject to flow
Liquefied Conditions liquefaction and the post-earthquake factor
of safety is significantly less than 1.0, large,
C3.3.4.1 Limit Equilibrium Analysis. rapid deformations and flow sliding can be
The post liquefaction stability is assessed as expected. If the factor of safety is only
follows (Fell et al, 2005, 2015): marginally less than 1.0, or marginally
(a) Determine the zones which have above 1.0, deformations may not be so large
liquefied (i.e. FS < 1.0) under the as to lose freeboard between the dam crest
earthquake loading using the and the reservoir level.
methods in Section C3.4.4. In considering these factors of safety
(b) Determine the liquefied shear account should be taken of the uncertainty in
strength (Su(LIQ)) for these zones
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 3.0 Assessment of Embankment Dams for Seismic Ground Motion
the assessment of the extent of liquefied liquefied strength. These are best accounted
zones and the liquefied strength. Perlea and for in a risk framework by assigning a
Beaty (2010) indicate that a factor of safety likelihood to a range of liquefied strength,
in excess of 1.2 to 1.5 may be required to and analysing for each liquefied strength.
achieve tolerable displacements.
C3.3.5 Advanced Numerical Methods for
Because of the uncertainty in what post Estimating Deformations During
liquefaction factor of safety will give and Post-Earthquake for Non-
tolerable displacements, this approach and Liquefied and Liquefied
the simplified method described in Section Conditions
C3.3.4.2 should only be used as described in The dynamic numerical codes used in
the Guideline. practice may be divided into two main
categories: total stress codes, and effective
C3.3.4.2 Simplified Deformation Analysis. stress codes (Zienkiewicz et al., 1986 and
An approximate estimate of the Finn, 1993). They are reviewed in
deformations can be obtained using the Zienkiewicz et al., (1986), Finn (1988,1993,
Khalili et al. (1996) method which is 2000), Marcuson et al. (1992), and Perlea
detailed Fell et al (2005, 2015). and Beaty (2010). A brief discussion of
some of the more frequently used codes

C4.3.4.3 Static Numerical Deformation
within each category is provided in the
Indicative estimates of deformations can be following sub-sections.
obtained by performing a static deformation C3.3.5.1 Total Stress Codes
numerical analysis which incorporates the The total stress codes, as can be inferred
earthquake induced pore pressures and the from the classification, are based on the total
liquefied strength of the liquefied soils stress concept and do not take account of
(Finn, 1993). The analysis is often pore pressures in the analysis. Therefore
performed in two stages. In the first stage, they should only be used in situations where
the numerical model is initialised to the pre- the seismically induced pore pressures are
earthquake conditions of the dam by negligible. The total stress codes may be

simulating the current in-situ stresses. Then, divided into two main categories: (1) codes
in the second stage, the earthquake induced based on the equivalent linear (EQL) method
pore pressures and residual strengths of the of analysis, and (2) fully non-linear codes.
liquefied soils are incorporated into the

model to simulate post-liquefaction (a) Equivalent linear analysis (EQL)

The earlier total stress codes were based on
This type of analysis is also referred to as the EQL method of analysis developed by
uncoupled deformation analysis and H.B. Seed and his colleagues in 1972. EQL
generally leads to conservative estimates of is essentially an elastic analysis and was
post liquefaction deformations, as it does not developed for approximating non-linear
allow for dissipation of earthquake induced behaviour of soils under cyclic loading.
pore pressures with time. More accurate Typical of the EQL codes used in practice
estimates of post liquefaction deformations are: SHAKE (Schnabel et al., 1972), QUAD-
can be obtained using fully and semi- 4 (Idriss et al., 1973), QUAD4M (Hudson et
coupled methods of analysis, as discussed in al 1994) and FLUSH (Lysmer et al., 1975).
the following sections. SHAKE is a one dimensional wave
propagation program and is used primarily
However for many projects the uncoupled for site response analysis. QUAD-4,
deformation analyses are sufficient. As for QUAD4M and FLUSH are two-dimensional
the more simplified methods, account should versions of SHAKE and are used for seismic
be taken of the uncertainty in the assessment response analysis of dams and
of the extent of liquefied zones and the embankments.
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 3.0 Assessment of Embankment Dams for Seismic Ground Motion
Given the elastic nature of the EQL analysis, related to recoverable (elastic) deformations
however, these codes cannot take account of and the residual pore pressures are related to
material yielding and material degradation non-recoverable (plastic) deformations. A
under cyclic loading. Therefore, they tend to major challenge in fully coupled codes is to
predict a stronger response than actually predict residual pore pressures. The residual
occurs. Also, they cannot predict the pore pressures, unlike the transient pore
permanent deformations directly. Indirect pressures, are persistent and cumulative and
estimates of permanent deformations can thus exert a major influence on the strength
however be obtained using the acceleration and stiffness of the soil skeleton. The
or stress data obtained from an EQL analysis transient pore pressures are cyclic in nature
and the semi-empirical methods proposed by and their net effect within one loading cycle
Newmark (1965) and / or Seed et al. (1973). is often equal to zero. An accurate prediction
of residual pore pressures requires an
(b) Fully non-linear analysis accurate prediction of plastic volumetric
deformations. In the fully coupled codes this
More accurate and reliable predictions of
is often achieved by utilizing elasto-plastic
permanent deformations can be obtained
models based on kinematic hardening theory
using the elasto-plastic non-linear codes.
of plasticity (utilizing multi-yield surfaces)
Typical of the elasto-plastic non-linear codes
or boundary surface theory with a hardening

used in the analysis of embankments are
law. These models are very complex and put
DIANA (Kawai, 1985), ANSYS (Swanson,
a heavy demand on computing time.
1992), FLAC (Cundull, 1993), etc. The
constitutive models used in these codes vary
Generally speaking, fully coupled prediction
from simple hysteretic non-linear models to of pore pressures under cyclic loading is
more complex elasto-kinematic hardening very complex and difficult. The validation
plasticity models. Compared to the EQL studies performed on a number of these
codes, the elasto-plastic non-linear codes are codes suggest that the quality of response
more complex and put heavier demand on predictions is strongly path dependent.
computing time. However, they provide When the loading paths are similar to the
more realistic analyses of embankments stress paths used in calibrating the models,

under earthquake loading, especially under the predictions are good. As the loading path
strong shakings. Critical assessments of non- deviates from the calibration path, the
linear elasto-plastic codes can be found in predictions become less reliable. Apart from
Marcuson et al. (1992) and Finn (1993, the numerical difficulties, part of this

2000). unreliability is also due to the poor or less

than satisfactory characterization of the soil
C3.3.5.2 Effective Stress Codes
properties required in the models. For
Effective stress codes allow modelling pore
instance, because of sampling problems, it is
pressure generation and dissipation in
often very difficult to accurately determine
materials susceptible to liquefaction and thus
volume change characteristics of loose sands
to obtain better estimates of permanent
as required by these models. In general, the
deformations under seismic loading. The
accuracy of pore pressure predictions in
effective stress codes may be divided into
fully coupled modes is highly dependent
three main categories: fully coupled, semi-
upon the quality of the input data.
coupled and uncoupled.
Fully coupled codes include DNAFLOW
(a) Fully coupled codes
(Prevost, 1981, 1988), DYNARD (Moriwaki
In the fully coupled codes, the soil is treated et al., 1988); PLAXIS (2012); ABAQUS
as a two-phase medium, consisting of soil (2012) and FLAC (Itasca, 2011).
and water phases. Two types of pore
As described in Perlea and Beaty (2010),
pressures are considered, transient and
FLAC contains a number of general purpose
residual. The transient pore pressures are
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 3.0 Assessment of Embankment Dams for Seismic Ground Motion
constitutive models but also provides for the subject to liquefaction, and where the extra
use of user-defined constitutive models. refinement of the analyses are warranted;
These include a bounding surface hypo- e.g. where the simplified methods are not
plasticity model (Wang and Makdesi, 1999) really applicable, and / or are resulting in
by AMEC Geomatrix Inc.; FLAC- marginal assessments on whether remedial
UBCSAND, (Beaty and Byrne, 2008) and works are required.
PM4Sand (Dafalais and Manzari, 2004;
Boulanger, 2010; Boulanger and Ziotpoulou, Any non-linear deformation analysis needs
2012; Ziotopoulou and Boulanger, 2012). to be documented in sufficient detail that it
PLAXIS and ABAQUS also offer user can be reasonably scrutinized. All input
defined constitutive models such as parameters, all numerical details (boundary
UBCSAND. conditions, damping parameters, etc), and
enough output plots to evaluate
(b) Semi-coupled codes reasonableness of initial conditions and
dynamic responses. These should be
Compared to the fully coupled codes, the reviewed by person’s expert in the practice.
semi-coupled codes are more robust and less
susceptible to numerical difficulties. Perlea and Beaty (2010) give an overview of
However they are theoretically less rigorous. USACE practice and examples of use of

In these codes empirical relationships, such many of the programs described above.
as those proposed by Martin et al. (1975) Stark et al (2012), Friesen and Balakrishnan
and Seed (1983), are used to relate cyclic (2012) give useful examples.
shear strains/stresses to pore pressures. The
empirical nature of the pore pressure C3.4 Assessment of the Effects of
generation in these codes generally puts less Liquefaction in Embankment Dams
restriction on the type of plasticity models and Their Foundations
used in the codes. The semi-coupled codes
C3.4.1 Overall Approach
are in general less complex and
No commentary.
computationally demanding. The parameters
they require are often routinely obtained in C3.4.2 Definitions and the Mechanics of

the laboratory or in the field. Liquefaction

It should be noted that there are no
Semi coupled codes include DESRA-2 (Lee
universally accepted definitions for
and Finn, 1978), DSAGE (Roth, 1985), and
liquefaction. Those provided in the

TARA-3 (Finn et al., 1986). FLAC (Cundall,

Guideline have fairly wide acceptance.
1993, Itasca, 2011) can also be used in this
mode as in the URS cycle weighted model The definitions for liquefaction, initial
described in Perlea and Beaty (2010) liquefaction, flow liquefaction and
temporary liquefaction are based on USNRC
C3.3.5.3 Summary (1985), Robertson and Fear (1995),
More advanced dynamic methods are
Robertson and Wride (1997), Yamamuro
potentially expensive and should only be
and Lade(1997).
done by very experienced persons who
understand the limitations of the analysis It is important to recognize that these
and the need to use well considered phenomena apply to monotonic (static) as
properties. The analysis methods are well as cyclic loading and are apparent in
controlled by the quality of data put into contractive soils (e.g. in loose fills) , both
them, the limitations of the methods cohesionless and those with very low
themselves and particularly of those doing plasticity.
the analysis.
Robertson and Fear (1995) and Robertson
They are in most cases only to be and Wride (1997) defined the cyclic
contemplated for dams and foundations liquefaction and cyclic mobility terms.
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 3.0 Assessment of Embankment Dams for Seismic Ground Motion
Idriss and Boulanger, (2008) defined cyclic embankment dams, aging effects are
softening. minimal, and corrections for age should not
be applied in calculating liquefaction
The mechanics of liquefaction are described resistance.
in USNRC (1985) and Fell et al (2005,
2015). However the description in Idriss and Andrus et al (2009) and Hayati and Andrus
Boulanger (2008) is more complete and up (2009) present methods for correcting for the
to date. age of the soils being assessed for
liquefaction. These are discussed further in
It is essential that those carrying out Section C3.4.4.3. These include data
liquefaction assessments read Idriss and showing liquefaction has occurred in
Boulanger (2008) because it will assist them Pleistocene sands.
understand the background to the simplified
procedure for liquefaction assessment and It is concluded that geological age alone
the way laboratory performance has been cannot be used as a means of screening for
used to develop the method alongside field liquefaction susceptibility. It may be taken
data. as additional information to be considered in
assessing the likelihood of liquefaction.
C3.4.3 Methods for Identifying Soils

which are Susceptible to (b) Soil gradation, plasticity, moisture
Liquefaction content

(a) Geology and age of the deposit. There are a number of methods available to
identify soils which are susceptible to
Idriss and Boulanger (2008) refer to Youd liquefaction based on the particle size
and Perkins (1978) and indicate that the soils gradation, plasticity and in-situ moisture
which are most susceptible to liquefaction content of the soil. These include:
are non-plastic or very low plasticity fill,
alluvial, fluvial, marine and deltaic soils. (i) Seed et al (2003) criteria
They indicate soils are most susceptible (ii) Robertson and Wride (1998) soil
when they are recently deposited, becoming behaviour type index Ic for Cone

more resistant to liquefaction as they Penetration Tests

become older. (iii)Boulanger and Idriss (2006)
(iv)Robertson (2010) for Cone
The most susceptible soils are Holocene or Penetration Tests

younger soils, Pleistocene soils are low or

very low susceptibility except for loess. It is recommended that at least two and
preferably more of the methods be used to
Youd et al (2001) discuss the effects of the assess the likelihood the soils are potentially
age of the deposit and conclude that for liquefiable. A suggested approach for doing
man-made structures such as fills and this is given in Table C3.1.
Table C3.1 Suggested method for assessing the likelihood that a soil is potentially

Result of assessment Likelihood this strata is potentially

Three methods indicate liquefiable Highly likely to certain
Two methods indicate liquefiable Likely
Only Seed et al (2003) indicates liquefiable, Likely to Possible
others do not Depending if Zone A or Zone B
Only Robertson and Wride (1998) indicates Unlikely
liquefiable, others do not.
Three methods indicate non liquefiable Liquefaction can be discounted
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 3.0 Assessment of Embankment Dams for Seismic Ground Motion
It should be noted that where soils plot in field case data which he indicates
Zone B of the Seed et al (2003) method they shows the Boulanger and Idriss
should be considered as potentially (2006) criteria are non-conservative,
susceptible to liquefaction and /or cyclic with case data showing liquefaction
softening. If their behaviour becomes critical of soils occurring with PI> 7 contrary
to the assessment of the safety of the dam to the Boulanger and Idriss (2006)
either the soil should be assumed to be criteria. Given these data it would
liquefiable or undisturbed samples should be appear that caution should be applied
subject to cyclic loading tests rather than in relying on the Boulanger and
relying on semi-empirical methods to predict Idriss (2006) criteria alone to identify
behaviour. soils which may liquefy.
(d) Fell (2012) notes that for one dam
The following points should be noted: where foundation liquefaction was
thoroughly assessed using SPT and
(a) The Seed and Idriss (1982) Chinese
CPT data the Robertson and Wride
Criteria method has been used
(1998) method identified a number of
extensively in Australia but often the
strata as non-liquefiable where the
last criteria that the natural moisture
Seed and Idriss (1982) and Seed et al
content is ≥ 0.9 times the liquid limit
(2003) methods both indicated the

has been ignored which mean that
soil was potentially liquefiable.
lower moisture content soils have
Given this caution should be applied
been incorrectly identified as
in relying only on the Robertson and
potentially liquefiable.
Wride (1988) or Robertson (2010)
(b) Seed et al (2003) and Boulanger and
CPT based methods and possibly
Idriss (2007), Idriss and Boulanger
other CPT and CPTU based methods
(2008) all indicate that the Chinese
Criteria (Seed and Idriss, 1982) are
(e) Particle size distribution alone is not
outdated and should not be used. It is
able to identify soils which are
recommended that this advice be
susceptible to liquefaction. Figures

11 and 12 in ANCOLD (1998) are

(c) Seed (2010) discusses the Boulanger
outdated and should not be relied
and Idriss (2006) method and plots
upon. Figure C3.3 is a better guide.


80 4
Percent passing (%)



Approximate bounds of fills
susceptible to static liquefaction
and flow sliding
0.001 0.01 0.1 1 10 100
Clay Silt Sand Gravel
Particle size (mm)
1 - Coarse grained coal mine waste (Dawson et al 1998; Taylor 1984; Bishop et al 1969; Hutchinson 1986)
2 - Loose silty sand fills, Hong Kong (upper and lower quartile of pre 1977 fills (HKIE 1998))
3 - Hydraulically placed mine tailings and fills in dam embankments (various published sources)
4 - Sensitive clays (indicative limits from: Lefebvre 1996; Bentley & Smalley 1984; Mitchell & Markell 1974; Hutchinson 1961, 1965)
5 - Sub-aqueous slopes, natural and fill slopes (Koppejan et al 1948; Kramer 1988; Sladen & Hewitt 1989; Cornforth et al 1974)

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 3.0 Assessment of Embankment Dams for Seismic Ground Motion
Figure C3.3: Particle size distributions of material types susceptible to liquefaction and flow
sliding (Hunter and Fell 2003)

C3.4.4 Methods for Assessing Whether opinion, often personalised and polarized”.
Liquefaction may occur This is reported in Finn et al (2010).

C3.4.4.1 Introduction Youd (2010) provided some personal

Whether soils are likely to liquefy under insights to the discussion.
seismic loading is almost always carried out
using the “simplified method” first Cox and Griffiths (2011) contributed by
developed by Seed and Idriss (1971). In the comparing the use of the Youd et al (2001),
ANCOLD (1998) guideline the method Seed et al (2003) and Idriss and Boulanger
referenced was Seed and De Alba (1986) (2008) approaches on three sites.
which was a refinement of the Seed and
As a result of these developments it has been
Idriss (1971) method. Since then there have
necessary in preparing these guidelines to
been further developments in the details of
review these documents and provide
the methods but the basic approach remains
guidance on which method or methods
the same as described in Section C3.4.4.2.
should be used, and details which need to be

Youd et al (2001) published a summary
report on the 1996 NCEER and 1998
Unlike the ANCOLD (1998) guideline
NCEER/NSF Workshops on valuation of
details of the suggested method are not
Liquefaction Resistance of Soils. This was a
included here. Those doing liquefaction
consensus of the 20 experts who attended the
assessments should instead work from the
Workshops and has been widely used for
paper, Monograph or Report so they have
assessing liquefaction since it was published,
available the background to the methods.
including by the Australian dam community.
It has been proposed by Finn et al (2010) that
Seed et al (2003) prepared a University of
a Workshop similar to the 1996 Workshop be
California at Berkeley Report which convened. Practitioners should seek that
presented updated methods based on

publication if and when it is produced. In the

enlarged databases of cases, and using a meantime they should keep up to date with
probabilistic approach.
the literature particularly that in high quality
Idriss and Boulanger (2008) from University refereed journals.

of California at Davis produced Monograph

Since the drafts of these guidelines were
MNO-12 for the Earthquake Engineering
prepared there have been a number of
Research Institute (EERI). This reviewed the
publications by Idriss, Boulanger, and others.
available literature and updated many aspects
These update and refine the methods but do
of the method.
not significantly alter the outcomes so no
Seed (2010) issued a University of California attempt has been made to review them here.
at Berkeley Geotechnical Report which was
C3.4.4.2 Outline of the “Simplified
critical of a number of aspects of Idriss and
Method” for Assessing
Boulanger (2008).
Idriss and Boulanger (2010) responded to The method requires the calculation or
this criticism and provided more details on estimation of two variables for evaluation of
the basis of their EERI Monograph. liquefaction resistance of soils:

As a result of the differences of opinion (a) The Cyclic Stress Ratio, CSR, which
EERI set up an ad hoc committee to review is a measure of the cyclic load
what they termed the “strong differences of applied to the soil by the earthquake.

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 3.0 Assessment of Embankment Dams for Seismic Ground Motion
(b) The Cycle Resistance Ratio, CRR, In this figure, (N1)60 is the SPT blow count
which is the capacity of the soil to normalised to an effective overburden
resist liquefaction. The CRR is pressure of 100 kPa, a hammer energy of
estimated from Standard Penetration 60%, borehole diameter, rod length and
Tests (SPT), Cone Penetration Tests sampling method.
(CPT) or, less frequently, the shear
wave velocity. (N1)60 is given by:
(c) If the CSR is greater than the CRR,
(N1)60 = Nm CN CE CB CR CS
liquefaction is likely to occur.
Where N m = measured standard penetration
The Cyclic Stress Ratio, CSR, is calculated resistance; CN = factor to normalize Nm to a
common reference effective overburden
CSR = (t av / s¢vo ) = 0.65(a max / g)(s vo / s¢vo )rd stress; CE = correction for hammer energy
ratio (ER); CB =correction factor for
borehole diameter; CR = correction factor for
Where a max = peak horizontal acceleration at rod length; CS = correction for samplers with
the ground surface generated by the or without liners.
earthquake; g = acceleration of gravity; svo
and s′vo are total and effective vertical To allow for fines content the (N1)60 values

overburden stresses, respectively and r d = are adjusted to (N1)60cs using:
stress reduction coefficient.
(N1)60CS = a + b(N1)60
The CRR for M7.5 earthquakes can be
estimated from the curves in Figure C3.4. It a and b vary with fines content.
is usually recommended that the SPT Clean
Sand Base Curve be used, with correction for
fines content to give (N1)60CS. There are
similar plots for CPT and shear wave

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 3.0 Assessment of Embankment Dams for Seismic Ground Motion

Figure C3.4 SPT Clean-sand Cyclic Resistance Ratio (CRR) curve for Magnitude 7.5
Earthquakes with data from liquefaction case histories (Youd et al., 2001, modified from Seed et
al., 1985).
Magnitude scaling factors MSF gently sloping ground and for depths less
than about 15 m.
The procedures outlined above give CRR7.5,
cyclic resistance ratios for M7.5 For assessments of liquefaction for
earthquakes. embankment dams, the confining stresses
may be higher than this and the dam will

Smaller magnitude earthquakes giving the impose static shear stresses, which may alter
same peak horizontal acceleration are less the liquefaction potential.
likely to initiate liquefaction, because the
earthquake will have fewer cycles of motion. These are allowed for by using the factors

Larger magnitude earthquakes are more Ks and Ka to allow for the high overburden
likely to initiate liquefaction. This is allowed stresses and static shear stress respectively.
for by using a magnitude scaling factor They are applied by extending the equation
(MSF) in the equation for factor of safety above to:
(FS) against liquefaction:
FS = (CRR7.5/CSR).MSF. Ks. Ka
There are three most commonly used
Where CSR = calculated cyclic stress ratio methods based on SPTs and CPTs:
generated by the earthquake shaking; and
CRR7.5 = cyclic resistance ratio for · Youd et al (2001)
magnitude 7.5 earthquakes. · Seed et al (2003), Cetin et al (2004)
and Moss et (2006)
Corrections for overburden stresses, static · Idriss and Boulanger (2008),
shear stresses Ks and Ka Boulanger and Idriss (2012)
The method of assessment of liquefaction Much of the discussion about the method
potential described above is for horizontal or relates to how the stress reduction factor rd,
magnitude scaling factor MSF, and
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 3.0 Assessment of Embankment Dams for Seismic Ground Motion
corrections for high overburden stresses and Overall it is concluded that:
static shear stresses Ks and Ka are treated.
There are also differences of view regarding 1. All three SPT based methods may be
the CRR plot. used and be expected to result in
similar factors of safety against
C3.4.4.3 Discussion of Differences between liquefaction in the upper 15 metres.
the available SPT and CPT Based 2. The Idriss and Boulanger (2008),
Methods Boulanger and Idriss (2012) SPT
Those who wish to understand the based method incorporates many
background to the methods should read the refinements developed since the
references in Section C3.4.4.1. Youd et al (2001) method was
published and its use is preferred.
Cox and Griffiths (2011) used the three SPT However where Ks values are critical
based methods to analyse three case studies. to the outcome both the Idriss and
They found that they did not give greatly Boulanger (2008) and Youd et al
differing factors of safety against (2001) / Hynes and Olsen (1999)
liquefaction in the upper 24m of the soil values should be used, and where the
profiles. They found that the Idriss and Idriss and Boulanger (2008) values
Boulanger (2008) method typically yields are controlling the outcome of the

the highest factor of safety at depth, assessment of the likelihood of
generally ranging from 5-15% higher than liquefaction expert advice should be
the others, but the differences were much sought.
greater at raw blow counts greater than 23.
3. As can be seen in Figures C3.5 and
This was attributed to Idriss and Boulanger C3.6 there is not a large difference
(2008) using higher Ks values but more between the NCEER/NSF
importantly higher CN values. These Workshops (Youd et al (2001) and
combined with the shape of the CRR curve Idriss and Boulanger (2004, 2008)
at high (N1)60CS to give the larger plots. Within the context of the
differences. uncertainty in ground motions the
differences are small and will make

They found that in the upper soil profile the

little difference in a risk based
individual inputs to the methods varied quite
analysis. The Cetin et al (2004) (Seed
a lot but the effects were often balanced by
et al 2003) curve plots to the right of
other factors. In particular they found that
the others. This apparent significant

the quite different rd values did not result in

difference between the three authors
greatly different factors of safety.
can partly be explained by the way in
This is consistent with the advice from Idriss which the database has been analysed
and Boulanger (2008) that because the case by Cetin et al (2004). However Idriss
histories used in the development of their and Seed (2010, 2012) present a
method were analysed using their methods discussion of the data points which
for accounting for depth effects, earthquake control the position of the curve.
magnitude and fines content, it is important They cast doubt on the validity of 8
that when using their method the same key points on the grounds of
methods as they used for these effects are incorrect assignation of liquefaction /
adopted. no liquefaction of four points and
significant numerical errors between
It is apparent that this would also apply if the rd values used to develop their
using the Seed et al (2003) method. In correlations and the rd values
particular values of rd , CN , and Ks should computed using their applicable
not be substituted from one method to equation for four others. In view of
another. this the Idriss and Boulanger (2004,

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 3.0 Assessment of Embankment Dams for Seismic Ground Motion
2008), Boulanger and Idriss (2012) marginal cases should be assumed
method is preferred. liquefiable, or laboratory tests and
4. There is less information available to numerical analysis of stresses should
make an assessment of CPT based be carried out to assess the effects.
methods. However on what is 6. The simplified methods are not
presented all three CPT based considered reliable below about 15 or
methods may be used for assessing 20 metres. Most of the uncertainty
factors of safety against liquefaction discussed above is below that depth.
in the upper 15 or 20 metres. For important decisions relating to
5. It is apparent there is considerable marginally liquefiable soils below
uncertainty in the estimation of Ka. It that depth expert advice should be
is suggested that Idriss and obtained.
Boulanger (2008) be used but


Figure C3.5 Curves relating cyclic resistance ratio (CRR) to (N1)60CS values for earthquake

magnitude Mw =7.5. (Idriss and Boulanger (2008).

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 3.0 Assessment of Embankment Dams for Seismic Ground Motion

Figure C3.6 Curves relating Cyclic Resistance Ratio (CRR) to (qc1N)cs values for earthquake
magnitude Mw =7.5. (Idriss and Boulanger (2008)).
7. There is considerable uncertainty in soils where SPT and CPT samplers are
many of the parameters used in these affected by the gravel particles.
methods. Users should not regard the
outcomes as clearly liquefiable or not C3.4.4.5 Allowance for the Age of the Soil
liquefiable for soils plotting near the Deposit
boundaries. This can be readily There are a number of authors who have
presented data to support the conclusion that

accounted for in a risk based

framework as these ANCOLD the age of the soil deposit has an effect on
Guidelines are recommending. the resistance of the soil to liquefaction.
These include Youd and Perkins (1978),
C3.4.4.4 Shear Wave Velocity ( VS) Based Seed (1979), Arango et al (2000), Leon et al

Methods (2006), Andrus et al (2009) and Hayati and

Youd et al (2001), Seed et al (2003) and Andrus (2009).
Idriss and Boulanger all refer to Andrus and
Stokoe (2000). Kayen et al (2012) have Andrus et al (2009) present an alternative
gathered more case data and have developed method based on measured to estimated
a probabilistic approach based on this. shear velocity VS. They include data which
shows how the shear wave velocity V S and
Idriss and Boulanger (2008) point out the CPT and SPT can give quite different
relative lack of sensitivity of VS to changes assessments of liquefaction potential in
in relative density. A relative density range Pleistocene and Tertiary sands.
of 30% to 80% results in about 7 x difference
in SPT “N” value, 3 x in CPT cone resistance It is suggested that these methods may be
and only 1.4 x in VS. As a result greater considered as additional information where
weight should be placed on SPT and CPT SPT and CPT based methods are showing
based methods. marginal potential for liquefaction in
Pleistocene and Tertiary sands. However the
This fundamental constraint means the methods are very approximate and not too
method is mostly restricted to use in gravelly much reliance should be made on them.
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 3.0 Assessment of Embankment Dams for Seismic Ground Motion
They should also be considered for a result considerable judgement is exercised
application in assessing the liquefaction by the authors in developing the methods,
potential of recently deposited mine tailings and the results are uncertain, with a wide
and dredged fills, as there is evidence the range of strengths possible.
resistance of these to liquefaction may be
lower than for the soils in the databases upon The issue is further complicated because
which the CRR are calculated. there are two schools of thought:

C3.4.4.6 Use of Becker Penetration Test (a) Those such as Seed et al (2003), Seed
In sandy gravel and gravelly sand soils the (2010) who believe that the liquefied
SPT sampler may hit gravel size particles residual strength Sr should be related
and register high blow counts even if the soil to the SPT (N1)60CS value directly
is not dense. In USA and Canada for these with no allowance for the effective
soils the Becker Penetration Test is overburden stress. This is termed a
sometimes used. “critical state” based method.
(b) Those such as Olsen and Stark (2002)
The Becker Penetration sampler is 168mm who believe the residual strength
diameter compared to 35mm diameter in the ratio Sr /s′vc should be used and
SPT sampler, and is driven by a down-hole related to the SPT (N1)60CS value.

diesel hammer. A modern improvement of That is the residual strength is
the equipment is described in Ghafghazi et al controlled by the effective
(2014a, b) and DeJong et al (2014, 2016a) overburden stress. Idriss and
Boulanger (2008) and Robertson
Some methods for interpreting the results are (2010) have developed such methods.
given in Sy and Campanella (1994), and Figure C3.7 shows the Idriss and
Ghafghazi et al (2014a, 2016b). Boulanger (2008) data and
C3.4.5 Methods for Assessing the Strength recommended curves. This is termed
of Liquefied Soils in the the “normalised strength ratio”
Embankment and Foundation approach.

C3.4.5.1 Introduction C3.4.5.2 Discussion

As discussed in Idriss and Boulanger (2008), There is no consensus on which of the two
Seed et al (2003) and Seed (2010) the approaches is best.
residual shear strength that the liquefied soil Idriss and Boulanger (2008) provide both

mobilizes in the field is affected by void methods without guidance on which to adopt
redistribution, particle intermixing and other but say they believe the residual strength
field mechanisms which are not replicated in ratio method provides a better representation
the laboratory. of the potential effects of void redistribution
As a result they conclude that the only than the liquefied residual strength, while it
practical methods for estimating the liquefied is recognised that neither correlation fully
strength Sr is by back analysis of case studies accounts for the numerous factors that
and relating these to SPT and CPT data. influence void redistribution processes.
However there are not a large number of
cases available, and often for these the SPT
and CPT data is variable and sometimes of
limited quality, and / or the geometry of the
embankment and the stratigraphy of the
foundation is complicated. Olsen and Stark
(2002) also describe some of these issues. As

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 3.0 Assessment of Embankment Dams for Seismic Ground Motion

Figure C3.7. Liquefied residual strength ratio Sr/s′vc versus SPT (N1)60CS value Idriss and
Boulanger, 2008).

Seed (2010) strongly criticises the Idriss and The Robertson (2010) CPT based method
Boulanger (2008) method, in particular the has the limiting value of liquefied undrained
“recommended curve for conditions where strength ratio is assumed to be 0.4 at Q1n,cs
void ratio redistribution effects are expected =70. This is based on the soils being dilative
to be negligible”. He argues that three of the at Q1n,cs > 70 so not being subject to flow

data points use incorrect and over stated liquefaction.

(N1)60CS ratios. He argues that the case data
does not support this curve. Baziar and Dobry (1995), Ishihara (1993)
and Cubrinovski and Ishihara (2000a, b)
Youd (2011) discusses this and says that have investigated the boundary of flow

Idriss and Boulanger unequivocally state that liquefaction conditions. Figure C3.8
the upper curve is based on laboratory data summarizes the outcomes. These support the
and was not guided by the points with use of larger liquefied strengths for soils with
incorrect strength in any way. He also says high (N1)60 values at least above (N1)60 = 15.
that a transformation from contractive to
dilative behaviour occurs at about a corrected Hence if the liquefied strength of such soils
SPT blow count of 15, and that should be is controlling the assessment of the safety of
accompanied by a major increase in shear the dam, and large costs are potentially
strength. He indicates he believes Idriss and involved in remedial works, it may be
Boulanger are “on the right track”. warranted to seek expert advice from one of
the experts in the field.

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 3.0 Assessment of Embankment Dams for Seismic Ground Motion

Figure C3.8. Comparison of flow liquefaction boundaries in terms of SPT (N1)60 for sands and
silty sands from monotonic laboratory undrained tests and earthquake triggered field case

(Hunter and Fell, 2003).
Given the lack of consensus on which the increase in svc′ from the berm. For high
approach to use (Critical state or normalised risk projects they suggest laboratory
strength ratio) it is suggested that both consolidation tests be carried out to confirm
methods be used for assessing the liquefied that it is parallel to the steady state line
residual strength, and the results accounted which is implicit in the assumption that S r
for assuming they are equally valid. /svc′ is constant. NSFW (1998) caution
against the use of Sr /s′vc ratios, particularly
Either the Idriss or Boulanger (2008), or
for clean sands. Given this and the
Seed and Harder (1990) / Seed (2010) or
uncertainty in which approach is most valid
Olsen and Stark (2002) methods may be

it would appear unwise to allow for the

used. These may be compared to the
strength increase from the increased effective
Robertson (2010) method.
stress of the berm and the Sr obtained
The majority of foundation liquefaction cases without the berm should be assumed to apply

in Australia will satisfy the requirements for after the berm is built, unless expert opinion
Idriss and Boulanger (2008) “recommended advises otherwise.
curve for where void redistribution effects
In all cases it should be recognised that the
could be significant” because the liquefiable
data indicates strength lower than the
layer is overlain by a lower permeability
suggested design values are possible, so
strata. In view of this and the reservations of
either a significant factor of safety is allowed
Seed (2010) it is suggested that only this
for or the uncertainty modelled in risk based
curve (not the upper one) be used unless
expert advice is sought.
C3.4.6 Methods for Assessing the Post-
An important issue is what liquefied residual
Earthquake Strength of Non-
strength to assume for design of stabilising
Liquefied Soils in the Embankment
berms for existing dams, or for the design of
and Foundation
new dams.
C3.4.6.1 Saturated Liquefiable Soils
Olsen and Stark (2002) indicate that in their
Soils which are potentially liquefiable but
view, at least for silty sands with > 12%
which have a factor of safety > 1.0 for the
fines, it is reasonable to allow for the
seismic load being considered will generate
increase in Sr which would be indicated by
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 3.0 Assessment of Embankment Dams for Seismic Ground Motion
pore pressures under the cyclic loading. are expected in the earthquake the
These pore pressures should be allowed for dense granular materials may have
in post-earthquake stability analyses. Idriss loosened and will have a strength
and Boulanger (2008) refer to Marcuson et al approaching the critical state strength,
(1990) for level ground conditions but warn rather than the peak strength. In
that pore pressures will be higher for practice this can be accommodated by
situations where static shear stresses are a small reduction from the expected
present. peak strengths.
The Marcuson et al (1990) data may be used C3.4.7 Site Investigations Requirements
to give a guide to what may happen. If these and Development of Geotechnical
become critical, and the project is Model of the Foundation
sufficiently important seek expert advice. The following are some suggestions
regarding site investigations of foundations
C3.4.6.2 Cyclic softening in Clays and of dams for assessing the likelihood and
Plastic Silts consequences of liquefaction.
Boulanger and Idriss (2006), Idriss and
Boulanger (2008) and Bray and Sancio (a) Review of available data.
(2006) describe how to estimate the effects

of cyclic softening of clays and plastic silts. Gather together all available data on the
The discussion relates to saturated soils with foundations including for existing dams,
over-consolidation ratios of 1 to 4. Such soils site investigations and laboratory testing
do occur in association with liquefiable soils, carried out prior to construction;
and may also occur as mine tailings. geological mapping, reports, and
photographs taken during construction;
The boundary between which this behaviour monitoring data; and site investigations
and liquefaction occurs is discussed in and laboratory testing carried out since
Section C3.4.3. the dam was constructed.

C3.4.6.3 Well Compacted Plastic and Non- (b) Preliminary geotechnical model and
Plastic Soils planning site investigations.

Fell et al (2005, 2015) suggest that:

Draw plans and sections of the available
(a) For clay soil zones in the data, taking account of the depositional
embankment and foundation assign environment of the foundation soils.

strength and pore pressures consistent

with the soil’s behaviour in static From this develop an interpretive
loading after being cracked and geotechnical model of the strata in the
disturbed by the earthquake. If the foundation, and summarize the properties
clay is contractive in nature, use in a form useful for liquefaction
undrained strengths. If it is dilative on assessment.
shearing, use effective stress Assess the adequacy of the available data
strengths c′, f′. Usually there will be and plan site investigations to supplement
some cracking and loosening, and if this data.
so fully softened strengths, would
apply (e.g. c′ = 0, f′ = f′peak). Some (c) Site investigation methods.
apply an arbitrary 10% or 15% loss of
strength. Foundations are best investigated by both
(b) For well compacted free draining boreholes in which SPT tests are carried
rockfill filters and dense sands and out on potentially liquefiable soils, and
gravels adopt the effective stress undisturbed thin wall tube samples taken
strengths c′, f′, with no change in the of other strata; and by Cone Penetration
pore pressures. If large deformations Tests (CPT, CPTU). This is desirable
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 3.0 Assessment of Embankment Dams for Seismic Ground Motion
because CPT are best able to detect shear wave velocity in addition to
layering in the strata, but SPT and boreholes with SPT.
laboratory tests are required to provide
data for some parameters, and to make The energy rating of the SPT hammer
available two methods for assessing the can be anything from around 55% to 90%
liquefaction potential and post- to the energy rating of the hammer(s)
earthquake liquefied strengths. used for the investigation should be
measured during the site investigations. It
Some of these (SPT and CPT) should be is not sufficient to use an energy rating
done adjacent each other for correlation for the hammer determined elsewhere
purposes. because it varies with the site conditions
and the drill equipment.
Boreholes should where practicable be
drilled using wash boring methods with Past practice has been to not measure the
casing and drilling mud support. To energy rating and assume 60%. This is
prevent “blow in” at the base of the hole not considered good practice and is
with resulting loosening of the soil in potentially conservative in many cases so
which SPT are carried out. the expenditure to measure the energy
rating is warranted.

Wash boring should not be used for holes
drilled through the core of a dam because For CPT and CPTU tests in layered soils
hydraulic fracture may be induced by the adjust the data for the effects of soft and
drill water or mud. If hollow flight augers stiff layers as detailed in Figure 7 of
are used (as will be necessary for holes Youd et al (2001).
drilled through an existing dam) take
measures to prevent “blowing” of the It should be noted that it is almost
strata at the base of the augers with impossible to take undisturbed samples
subsequent reduction in SPT blow of non-plastic soils such as those
counts. potentially liquefiable soils in the
foundations or mine tailings. The
Sonic drilling may be used but may result sampling process disturbs and potentially

in densification of the soil below the compresses the soil. Even specialised
casing. samplers are likely to result in
disturbance. It is for this reason most
Take samples from all SPT tests in analyses are based on in-situ testing.

potentially liquefiable strata and test for

moisture content and fines content (% (d) Geotechnical model
passing 0.075mm sieve). For
representative samples test for Atterberg Draw plans and sections of the available
limits. data including the data from the new
investigations, taking account of the
Record the level of the water table during depositional environment of the
drilling, and preferably install foundation soils.
piezometers to monitor groundwater
pressures at the time of the investigations From this develop an updated interpretive
and as they fluctuate with river flows. geotechnical model of the strata in the
foundation, and summarize the properties
For SPT tests in gravelly soils take blow in a form useful for liquefaction
counts for each 75mm as an aid to assessment.
detecting interference by gravel particles
and allow adjustment to allow for this. Take particular attention to the continuity
For gravelly soils consider the use of and aerial extent of potentially liquefiable

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 3.0 Assessment of Embankment Dams for Seismic Ground Motion
(e) Laboratory cyclic shear testing. close to fully saturated that they may
For many of the dynamic numerical (ii) Soils in the capillary zone above the
analyses methods cyclic simple shear and phreatic surface may be sufficiently
cyclic triaxial tests will be required to saturated to liquefy. However the
provide inputs into the models. Expert height of capillary rise is not likely to
advice should be obtained on the nature be large for sandy soils.
and detailed requirements for such tests. (iii)As noted in Section 3.1, for tailings
dams using upstream or centreline
C3.4.8 Liquefaction Analysis
construction, there is a potential for
The analysis of liquefaction potential is best
liquefaction of loose to medium
carried out as follows:
dense partly saturated tailings where
(a) Assemble the SPT and CPT /CPTU perched water tables or nearly
data for potentially liquefiable strata saturated zones may exist above the
into spreadsheets. In so doing assess measured phreatic surface. This may
the data for quality and lead to a series of persistent saturated
inconsistencies and remove or “tag” layers within the tailings.
as unreliable poor quality or (iv)Hossain (2010), Yang et al (2004),
Nakazawa et al (2004) and

inconsistent data.
(b) For SPT include all “N” in the Tsukamoto et al (2002) present
spreadsheet. For CPT / CPTU enter information which shows that the
data at 50 mm intervals as recorded. cyclic resistance of partially saturated
(c) Analyse the data to determine if the sands is higher than saturated sand,
soils are potentially liquefiable using and that these can be related to the
the criteria described above. Exclude compressional (“P” wave) velocity.
from further considerations soils
which are not liquefiable, or which C3.4.9 Assessment of the Likelihood and
the Effects of Cracking of
are above the water table, and will
not be saturated under any condition. Embankment Dams Induced by
Seismic Ground Motions

(d) For the potentially liquefiable strata,

Fell et al (2008), Section 5.5 describes a
asses the factor of safety against
method for estimating the likelihood of
liquefaction for a number of peak
ground accelerations within the range transverse cracks in the embankment caused
by earthquake. It relies on the damage

to be assessed for the site.

classification method of Pells and Fell (2002,
(e) From these data, plot plans and
sections showing contours of the
Annual Exceedance Probability Fell et al (2008) uses these data to estimate
(AEP) of liquefiable zones. the likelihood of failure of the dam given the
When assessing whether the soil may be
saturated a cautious approach should be These methods are empirical, and very
adopted taking into account the following: approximate.
(i) For some dams the tailwater level C3.5 Methods for Upgrading
may vary throughout the year and Embankment Dams for Seismic
during floods. As a result some soils Ground Motions
in the foundation and in the
embankment which are potentially C3.5.1 General approach
liquefiable may be below the water
table part of the time. It may be No commentary
appropriate to assume the soils in the
water table range are sufficiently
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 3.0 Assessment of Embankment Dams for Seismic Ground Motion
C3.5.2 Embankment Dams not Subject to surrounding soil should be taken as
Liquefaction the liquefied strength unless it can be
demonstrated that the soil between
No commentary. the columns is sufficiently permeable
to allow dissipation of pore pressures
C3.5.3 Embankment Dams Subject to built up during the cyclic loading.
Liquefaction This can be assessed by the method
of Seed and Booker (1977). Idriss and
Mitchell (2008), Idriss and Boulanger (2008)
Boulanger (2008) indicate that
and Seed et al (2003) discuss the various
method has been refined by Onoue
ground improvement methods and their
(1988), Iai and Koizumi (1986) and
limitations. These are summarized in Table
Pestana et al (2000), but they caution
that uncertainties in the hydraulic
The most widely used methods in Australia conductivities of the surrounding soil
have been “stone” columns, deep soil (particularly as it is disturbed by
mixing, and removal and replacement. construction of the columns) and the
columns themselves make it difficult
“Stone” columns are the most common but to have confidence of efficient
there are detailed design matters which need drainage except in relatively high

to be considered. These include: permeability soils.
(c) Seed et al (2003) discuss the effect of
(a) If the columns are situated on the the stiffer stone columns attracting
downstream side and there is no the cyclic shear stresses and
cutoff through the alluvium upon potentially reducing the build-up of
which the dam is constructed, the pore pressure in the liquefiable soil.
columns may acts as drains for They caution against relying upon
seepage water. In this case they this because the columns still flex if
should be backfilled with sandy they are longer than three times their
gravel designed to act as a filter to the diameter.
surrounding soil, or backward erosion

may initiate into the columns. This Some examples of seismic upgrades of dams
can create problems with the are given in Davidson et al. (2003), Toose et
construction of the columns as the al. (2007), Mejia (2005), Dise et al. (2004)
backfill is low permeability. For and Luehring et al. (2001)

some projects the backfill has been

designed as a “some erosion” filter,
using the Foster and Fell (2001)
(b) The composite strength of the
columns and the surrounding soil is
estimated allowing for the relative
areas of the columns and the
surrounding soil. The strength of the

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 3.0 Assessment of Embankment Dams for Seismic Ground Motion
Table C 3.2. Summary of ground improvement methods and their application for remedial
works for liquefiable foundations (Fell et al, 2015)

Method Method of Soils for which Limitations and Comments

Improvement it the method is
Vibroflotation Densification, Sand, sand with Ineffective if silt content > 20%
increased lateral some silt, approx. May not penetrate gravelly
stresses. gravelly sand strata.
Vibro- Densification, Sand, silt, clay Ineffective densification if silt content
replacement increased lateral > 20% approx. May not penetrate
e.g. stone stresses, gravelly strata. Columns may be
columns reinforcement, constructed by driving casing,
increased removing soil from within casing, and
drainage. compacting “stone” as casing is
withdrawn. Treatment must penetrate
through liquefiable soil or treatment
will be ineffective at the base.
Dynamic Densification, Sand, silty sand Ineffective if silt content > 20%

compaction increased lateral approx. Effective only in the upper
stresses. 10m. Quality control difficult.
Vibration of adjacent structures.
Compaction Densification, Sand, silt, clay Ineffective at depths < 6m. Augers
grouting increased lateral must penetrate through liquefiable
stresses, soil or treatment will be ineffective at
reinforcement. the base. Quality control may be
Deep soil Reinforcement, Sand, silt, clay Augers must penetrate through
mixing reduce earthquake liquefiable soil or treatment will be
induced strains ineffective at the base. Quality control

and pore may be difficult. Soil-cement

pressures, elements may be brittle.
Jet grouting Reinforcement, Sand, silt, clay Uses very high pressures so should
reduce earthquake not be used where hydraulic fracture

induced strains may damage the embankment core.

and pore Difficult to control diameter of
pressures, treatment.
Vertical Drainage Sand, sand with Ineffective in silty sands and silts
drainage some silt
Permeation Reinforcement Sand, gravel, Particulate grouts e.g. microfine
grouting sandy gravel, cement will not penetrate fine sands
gravelly sand because the sand acts as a filter.
Chemical grouts expensive and some
toxic. Quality control difficult.
Removal and Removes All soils Dewatering, stability during
replacement liquefiable soil construction may be problems. Gives
high degree of confidence in final
C3.6 Flood Retarding Basins

No commentary.

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 4.0 Assessment of Concrete Dams for Seismic Ground Motion

C4 ASSESSMENT OF CONCRETE The M7.6 Chi Chi Earthquake in Taiwan

DAMS FOR SEISMIC GROUND (1999) caused severe damage to the Shih
Kang Dam (a gated concrete gravity
structure) due to a fault rupture under the
C4.1 Effect of Earthquakes on Concrete dam causing a 9m vertical differential
Dams displacement of the dam. The rest of the dam
which was subjected to just the earthquake
C4.1.1 General shaking, performed well. Notwithstanding,
It was stated in ANCOLD (1998) the large differential displacement of the
“Guidelines for Design of Dams for dam, there was no uncontrolled loss of
Earthquake” and also ANCOLD (2013) storage so technically, the dam did not fail.
“Guidelines on Design Criteria for Concrete
Gravity Dams” but is worth repeating here, Major fault displacements directly under a
that concrete dams do fail. However, there concrete dam are very difficult to design for.
are no recorded cases where concrete dams Consequently, it is not included in the scope
have completely failed under earthquake of these guidelines. It is covered by ICOLD
loading, with the sudden release of the (1998).
reservoir. However, a number of concrete

ANCOLD(1998) included a section
dams have suffered very serious damage, “Analysis and Design of Concrete Dams”.
including cracking right through the However, that section was limited in that it
concrete. The two best known dams are effectively covered only concrete gravity
Koyna Dam (a 103m high gravity dam in dams. Since that time, there have been a
India) which suffered major cracking through number of earthquake analyses carried out in
the upper part of the dam when subjected to a Australia, of other types of concrete dams
M6.5 earthquake in 1967 and Sefid Rud Dam (e.g. Bendora Dam – a double curvature arch
(a 106m high massive buttress dam in Iran) dam in the ACT; Carcoar Dam – a double
which suffered major cracking in the curvature arch dam in NSW; and Oberon
buttresses when subjected to a M7.3-7.7 Dam – a buttress dam in NSW; Moogerah
earthquake in 1990. A number of other

Dam – a double curvature arch dam in

concrete dams of all types have also suffered Queensland; Mt. Bold Dam-a gravity-arch
damage under earthquake loading. dam in SA, see McKay and Lopez (2015);
Generally, the magnitude of the earthquake Clover and Junction Dams-slab and buttress
causing damage has been M5.3 or greater.

dams in Victoria). As there were no relevant

Damage has ranged from displaced copings local guidelines to cover these dams, resort
and minor cracking to major cracks with had to be made to reference material from
consequent leakage from the dam. Hansen overseas organisations (e.g. Federal Energy
and Nuss (2011) and Wieland and Regulatory Commission (FERC) and the US
Chen(2010) give information on the Army Corps of Engineers (USACE) in the
experiences of concrete dams subjected to USA).
earthquake loadings. Arch dams have tended
to perform well under earthquake loading. ANCOLD (1998) also had a number of other
The 113m high Pacoima Dam (a double shortcomings in today’s (2016) context:
curvature arch dam) in California was
subjected to a ground shaking of · They were oriented to safety reviews and
approximately 0.5g (horizontal and vertical) to standards based acceptance criteria;
during the Northridge Earthquake in 1994. · Acceptance criteria only given for gravity
The base accelerations were strongly dams (see discussion above);
magnified to 2g horizontal and 1.4g vertical · There was a detailed description given
on the top of the left abutment: the dam for a popular simplified methodology for
suffered only minor damage. analysing 2d gravity dams (fenves &
Chopra (1986) method) which was
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 4.0 Assessment of Concrete Dams for Seismic Ground Motion

probably unnecessary as documentation · Bulletin 120 (ICOLD, 2001): Design

for this methodology is readily available features of dams to effectively resist
via the internet; seismic ground motion
· little information was given regarding · Bulletin 123 (ICOLD, 2002): Earthquake
analysis required for risk assessments design and evaluation of structures
(i.e. analysis required to help estimate appurtenant to dams
conditional probabilities of failure of · Bulletin 155 (ICOLD, 2013): Guidelines
concrete dams); for use of numerical models in dam
· advances in computer software for engineering
undertaking sophisticated analysis of all
types of concrete dams both in 2D and The above Bulletins should generally be
3D space since ANCOLD (1998) was considered as background information.
published; and,
C4.2 Defensive Design Principles for
· recent publication of the updated
Concrete Dams
ANCOLD “Guidelines on Design No commentary.
Criteria for Concrete Gravity Dams”,
(ANCOLD, 2013). C4.3 General Principles of Analysis for

Seismic Ground Motions
The Chapter of the Guidelines and this
Commentary relating to concrete dams C4.3.1 Gravity Dams
attempts to resolve the above shortcomings ANCOLD (2013) includes a short section on
by: the design and analysis of gravity dams for
earthquake loading. That section is very
· emphasising the strategy of general in its nature and points to these
increasing the complexity of analysis Guidelines for information regarding gravity
only as the need requires; dams subjected to earthquake loading.
· ensuring that the level of
sophistication of the analysis is Concrete gravity dams including roller
consistent with the available compacted concrete (RCC) gravity dams can

information; generally by analysed in the first instance, as

· providing more information on the two dimensional structures – see Section
analysis of arch and buttress dams; C4.3.4 below for exceptions. This means
· providing more guidance on non- that relatively simplistic analysis methods

linear methods in recognition of more can be used to estimate the effects of

sophisticated software available. earthquake loadings on these dams. These
include the methodology of Fenves and
C4.1.2 ICOLD Chopra (1986) which will be discussed in
It should be noted that ICOLD has a number Section C.4.4.3.
of Guidelines relating to earthquake analysis
and design of concrete dams. A number of C4.3.2 Arch Dams
relevant Guidelines published this Arch dams include single curvature arch
millennium are given below: dams (most of which were built in this
country, from the end of the 19th century to
· Bulletin 62A (ICOLD, 2016b): the beginning of the 20th century), double
Inspection of dams following curvature arch dams (which can range from
earthquakes - guidelines thin to thick cross sections) and arch gravity
· Bulletin 148 (ICOLD, 2013): Selecting dams. The principles discussed in this
seismic parameters for large dams- chapter also apply to gravity dams which are
guidelines in relatively narrow, steep sided valleys and
are therefore should be considered as three
dimensional structures. Due to the relatively
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 4.0 Assessment of Concrete Dams for Seismic Ground Motion

flexible nature of arch dams (especially thin stresses that could exacerbate cracking when
double curvature arch dams) and their earthquake stresses are added.
inherent 3D nature, it can be necessary to
consider factors not applicable to gravity C4.3.4 Other Types
dams or possibly, buttress dams. These Arch gravity dams and gravity dams built in
include: water compressibility; dam-water narrow, steep sided valleys will be generally
interaction; wave absorption at the storage reviewed according to the concrete gravity
boundary and dam/foundation interaction. It guidelines. However, 3D analysis will
is usual practice to assume that the generally be required in addition to any
foundation rock is massless and simplified 2D analysis to ensure that
compressibility of the water and dam-water undesirable stability conditions/stress
interaction ignored if the fundamental regimes are not developed in the dam.
frequency of the storage is greater than that
C4.4 Seismic Structural Analysis of
of the dam and foundations alone. Duron et
Concrete Dams
al (1994) give a procedure for calculating the
fundamental frequency of the storage. C4.4.1 Material Properties
C4.3.3 Buttress Dams C4.4.1.1 Concrete

Buttress dams can range from mass buttress The compressive and tensile strengths of
dams which are akin to concrete gravity concrete increase as the rate of loading
dams to thin buttress dams which are increases. The dynamic compressive and
essentially reinforced concrete structures.
tensile strengths of concrete can therefore be
expected to be greater than the static
Slab and buttress type dams can be sensitive strengths. As dynamic compressive stresses
to cross valley shaking. The reinforced
are rarely of concern, the allowable
concrete slabs and the buttress corbels compressive stress for static loading can be
supporting them have to be able to handle
used also for dynamic loading. Raphael
upstream/downstream shaking. However, (1984) states that the apparent tensile
the relatively thin buttresses and the strutting
strength of concrete under seismic loading

system between the buttresses have to be which should be used with linear finite
able to handle the cross valley shaking as
element analyses is given by:
well as upstream/downstream shaking. The
interface between the slabs and their supports fr= 0.65 fc2

has to be able to safely transmit the shear where fc is the concrete compressive strength
forces between them. in MPa and fr is the apparent seismic tensile
strength in MPa. Values given by this
Applying concrete buttressing to the
formula are some 50% greater than the
downstream face of an existing concrete
apparent tensile strength for static loading.
gravity dam is a common form of
Raphael suggests that fr is twice the splitting
raising/strengthening those dams (e.g.
strength of the concrete under static loading.
Burrinjuck Dam in NSW, the spillway at
Clough and Ghanaat (1993) suggest that the
Hinze Dam in Queensland, Wellington Dam apparent dynamic tensile strength is about
in WA). This buttressing can be either
25% greater than the measured static value
continuous (Hinze and Wellington Dams) or which gives apparent tensile strength about
intermittent (Burrinjuck Dam). Analysis of
20% of the standard compressive strength.
these dams for earthquake loadings (as well They further suggest that there may be a 15
as static loads applied post buttressing) has to
to 20% loss of strength across lift joints.
consider the stress state of the dam prior to These figures may be even lower for poorly
applying the buttressing (i.e. how composite
constructed or defective lift surfaces.
action is set up between the original and new However, the peak dynamic tensile stresses
concrete) as well as any other locked in
only exist during a fraction of a response
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 4.0 Assessment of Concrete Dams for Seismic Ground Motion

cycle. Even though these peak stresses may the tensile strength of the rock may be
greatly exceed the tensile strength of the included for small concrete structures less
concrete, any cracking that might be initiated than about 5m high
will not have time to fully develop. It is well
recognised that a single spike of excessive C4.4.2 Loads
localised tension should not be taken to
represent dam failure.
Static loads applied to a dam for seismic and
In consideration of the above however, these post-earthquake analyses for all types of
guidelines recommend that for sound lift concrete dams will generally be the same as
surfaces, the apparent tensile strength to be those defined in the ANCOLD (2013)
used is 16% of the standard compressive
C4.4.2.2 Uplift
No commentary.
For dynamic modulus of elasticity, Clough C4.4.2.3 Silt
and Ghanaat (1993) suggest a value 25% No commentary.
greater than the static value and these
guidelines recommend this be adopted. For C4.4.2.4Seismic

existing dams, the elastic modulus of the
concrete mass may be determined using General
geophysical means (e.g. derived from Earthquake ground motion occurs in all
measured shear wave velocity). Values directions. Consequently, it is recorded in 3
obtained should be compared with static and orthogonal directions: 2 horizontal and one
dynamic small sample laboratory test values vertical. For 2D linear elastic models,
for credibility. sometimes only a single response spectrum
(see below for a discussion on response
C4.4.1.2 Foundations spectra) is available. In this case, it is
The Young’s modulus of deformation for a necessary to consider what vertical
jointed rock mass can be estimated from the earthquake motion might be concurrent with

Geological Strength Index (GSI) using the the horizontal shaking. In the past, the
method proposed by Hoek et al. (1998) and vertical acceleration used in analysis has
Hoek and Brown (1997. Fell et al (2015) been taken as a proportion (e.g. half to 2/3)
describe these and the method by Douglas of the peak ground acceleration. USACE

(2003). The dynamic modulus may be higher (2005) gives the following multipliers for
than the static modulus as discussed in vertical spectral acceleration to horizontal
Clough and Ghanaat (1993) and Scott and spectral acceleration depending on the
Von Thun (1993) and may, as for concrete, distance of the site from the earthquake
be obtained by geophysical means, or by generating fault:
relation to the static modulus.
· <= 10km distance: Sa| V/Sa|H = 1.00
A dam's foundations will normally contain · 25 km distance: Sa|V/Sa|H = 0.84
joints, shears, and bedding. Consequently, it · >= 40km distance: Sa| V/Sa|H = 0.67
will not be possible to transmit tensile stress
within the foundations and the allowable It is ANCOLDs preferred approach to
tensile strength for the foundations will develop a vertical to horizontal ratio specific
therefore normally be assumed to be zero. for the site as described in Section C2.8.
However, if extensive site investigation and
strength testing is able to prove that the
Inertial loads are those loads caused by
foundations for a particular dam site are
earthquake accelerations acting on the mass
massive rock, e.g. massive sandstone, are
of the dam. The acceleration at a particular
capable of transferring tensile stresses, then
level in the dam will generally be an
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 4.0 Assessment of Concrete Dams for Seismic Ground Motion

amplification of the ground acceleration due slope of the dam’s upstream face and the
to the flexibility of the dam and its response conditions at the bottom of the storage. For
to the earthquake shaking. This acceleration example, the hydrodynamic pressure
is applied to the incremental mass of the dam distribution of Fenves and Chopra (1986)
and added water mass (where added mass is takes into account the depth of the storage
used as against coupled hydrodynamic with respect to the height of the dam and the
pressures - see Hydrodynamic Loads) to energy absorbing characteristics of the
obtain the inertial loads. For superposition storage bottom.
methods, there will be a set of inertia loads
for each mode of vibration. USBR (2011) presents a discussion on a
number of methods for estimating
Hydrodynamic hydrodynamic pressures. The reference is
specifically for spillway gates but is relevant
Gravity and Buttress Dams: for concrete dams as well.
For gravity and buttress dams which can be
considered sufficiently rigid or for initial Arch Dams:
analysis of more flexible dams, the water can For thin arch dams, the assumption of
be assumed to be incompressible. The incompressible water and rigid dam do not

Westergaard (1933) pressure distribution is necessarily hold true. If the fundamental
therefore valid. This pressure distribution is frequency of the dam (without water) is
considered equivalent to a parabolic added approximately equal to the fundamental
mass of water fixed to the face of the dam. resonant frequency of the water in the
The incremental mass of water ( ) in Kg is storage, then the dam/storage interaction
given by: should be considered in more detail. The
fundamental resonant frequency ( ) in Hz
= ( − )A of the storage is given by:
where: =
H = depth of water (m)

ρ = density of water (Kg/m3) = ℎ ℎ (m)
y = (m/s)
= height of incremental added mass of water,
(m) = above the base of the dam

A = For the case where the dam is flexible and

incremental surface area on face of dam the water is compressible, finite element
(m2) modelling should include the storage
modelled using fluid elements, in addition to
the dam. A valuable discussion on this matter
The above only applies for a dam having a is given in FEMA (2014).
vertical upstream face and a straight axis.
For a dam having a sloping or curved C4.4.2.5 Dynamic Earth Pressures
upstream face, the acceleration normal to the No commentary.
face is used when calculating the
hydrodynamic force. In this case Zanger C4.4.2.6 Load Combinations
(1952) should be used for rigid dams. No commentary.

The Westergaard pressure distribution has C4.4.3 Methods Available and When to
been the standard method for many years. Use Them
However, consideration should also be given Earthquake analysis of a concrete dam may
to alternative hydrodynamic pressure be carried out in the frequency domain or in
distributions that are more relevant to the the time domain. In the former, a response

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 4.0 Assessment of Concrete Dams for Seismic Ground Motion

spectrum is used which has been derived for histories or synthetic time-histories.
the dam site from available seismic data. In Recorded time-histories have to be derived
the latter, a time-history for ground from earthquakes having a peak ground
acceleration or velocity is used that again, is acceleration (PGA) relevant to the annual
site specific. exceedance probability (AEP) earthquake
used to analyse the dam. For extreme
A response spectrum is usually prepared by a earthquakes (e.g. the SEE), recorded time-
specialist Seismologist as described in histories will generally not be available.
Section 2.9 . The response spectrum is However, synthetic time-histories can be
normally a plot of spectral acceleration generated from recorded time-histories
versus natural frequency or natural period of including ones from overseas earthquakes.
a single degree of freedom oscillator (e.g. a As described in Section 2.9 two methods for
lumped mass on top of a vertical, massless generating these synthetic time-histories can
cantilever). Sometimes however, it is be used: amplitude scaling or spectral
presented as a tri-partite plot giving spectral matching. Advice should be obtained from a
acceleration, spectral velocity and spectral specialist Seismologist regarding the method
displacement all plotted against frequency or appropriate for a particular dam.
period. The response spectrum gives an

indication as to how ground accelerations are There are a number of methods available to
magnified through the height of a dam. For a analyse concrete dams ranging in increasing
more detailed description of response spectra level of sophistication and complexity. These
and their theoretical background, the reader are detailed in Table C4-1 of the Guideline,
is referred to Clough and Jenzien (2003). which is taken from ICOLD (2013):
Time-histories of ground acceleration or
ground velocity can be either recorded time-

Table C4.1 Methods of analysis of concrete gravity dams for seismic ground motions.
(ICOLD, 2013)

Initial static conditions and loads

Initial state of the dam (cracking , joints)…
Static load conditions including uplift pressure
Selection of design earthquake ground motion characteristics

Level of Seismic Analysis Input Output

Level 0: preliminary PGA or effective Damage indices
screening acceleration, PGA, C, PSa(ξ,Ti)
Relative evaluation of seismic coefficient (maps),
seismic smooth design spectra
vulnerability of a portfolio
of dams
Level 1: pseudo-static Effective acceleration, Pseudo-static max.
analysis with constant seismic stress and velocity
seismic coefficients coefficient (maps),
Equivalent static forces, Westergaard added mass
rigid body equilibrium - 2D
gravity method

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 4.0 Assessment of Concrete Dams for Seismic Ground Motion

Level of Seismic Analysis Input Output

Level IIa - Pseudo-dynamic Smooth design spectra, SRSS (CQC) of max.
analysis hydrodynamic interaction, stress and velocity
(Chopra 1988) foundation interaction
Standard cross-section
gravity dams
equivalent static forces
rigid body equilibrium
2D gravity method
Level IIb - Linear response Smooth design spectra, SRSS (CQC) of max.
spectra reservoir stress and velocity
(modal) analysis added mass, Westergaard or
Classical modal analysis - fluid
Arbitrary cross section elements, Foundation model
Level IIIa - Linear time- Spectrum compatible Time-history
history analysis - accelerograms, analytical envelopes of stress,
frequency domain reservoir compressible velocity, acceleration,
Linear elastic FE analysis model, analytical foundation displacement

EAGD84, EACD-3D model with visco-elastic
half space
Level IIIb - Linear time- Spectrum compatible Time-history
history analysis accelerograms, added envelopes of stress,
- time domain Westergaard mass (or added velocity, acceleration,
Linear elastic FE analysis 2- mass using an alternative displacement
D, 3-D) method e.g. Fenves &
Chopra, 1986) for reservoir
or fluid elements,
foundation model

Level IVa - Non-linear Spectrum compatible Cracking response,

time-history accelerograms, added stability of cracked
analysis - Time domain Westergaard mass (or added components
Finite element analysis mass using an alternative
fracture analysis (cracking) method e.g. Fenves &

Level IVb - Non-linear hybrid Chopra, 1986) for reservoir

time-frequency or fluid elements,
domain (HFTD methods) - foundation model,
EAGD slide fracture material properties
Solution of non-linear (Kic, Kiic, Gf)
dependent equations of

These ANCOLD Guidelines are generally C4.4.4 Analysis in the Frequency Domain
consistent with this table by ICOLD. Further
details on numerical methods appropriate to Linear Elastic
the earthquake analysis of concrete dams can
be found in the ICOLD Guidelines as well as (a) The simplest method (item (i) above)
USBR (2006) and FEMA (2014). for undertaking a pseudo-static
stability analysis of a concrete gravity
dam is to use the method of Fenves
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 4.0 Assessment of Concrete Dams for Seismic Ground Motion

and Chopra (1986) with a site heel/face for either flood or

specific response spectrum for the earthquake cases, especially if
size (AEP/PGA) earthquake being cohesive strength is being relied
examined. In general terms, the upon.
natural frequency or period of the (b) Most commonly available finite
dam is calculated for a number of element analysis (FEA) software
modes of vibration. The spectral packages appropriate for concrete
acceleration is then determined for dam analysis include the ability to
these frequencies or periods, from the determine the natural frequencies and
design response spectrum for the dam mode shapes for a structure and to
site. Detailed documentation on the carry out a response spectrum
method is readily available via the analysis such that stresses induced by
internet. The method can be worked earthquake loading can be
through on a spreadsheet and assumes determined.
a triangular dam on a rigid foundation (c) The most recent review of pseudo
to calculate the fundamental natural dynamic analyses using finite
period. This natural period is then element methods is given in
modified using supplied charts to FERC (2016). This reference inter

estimate the effective fundamental aila, compares Westergaard’s Method
natural period and damping factor for hydrodynamic pressures with
accounting for the elastic modulus of other researchers.
the concrete in the dam relative to
that of the rock foundation, the depth The extreme response values for the various
of water against the dam and the modes of vibration are combined using the
amount of seismic wave reflection at square root, sum of the squares method
the bottom of the storage. The (SRSS) or complete quadratic combination
generalised mass of the dam and the method (CQC) combination. In turn, they are
earthquake force coefficient (which further combined using the appropriate
allows the multi degree of freedom combination method to include the effects of

dam to equate to a single degree of all three components of the earthquake

freedom oscillator) are then ground motion. The resulting overall
calculated. The hydrodynamic effect dynamic response values obtained will have
of the water is added as a series of no sign and may be considered to be either

lumped masses using Westergaard’s positive or negative. The SRSS method

Theory to the lumped masses of the usually provides a conservative estimate of
dam (see Sub-section C4.4.2.4). the maximum response when the vibration
Inertial forces at the levels of the periods of the dam structure are well
lumped masses are then calculated, separated. However, due to the method
with an adjustment that allows for ignoring the correlation between the adjacent
higher modes of vibration. These modes, the total response for the closely
forces can then be added to the static spaced vibration periods is underestimated.
forces on the dam, in a normal 2D In that case, it is better to use the CQC
cantilever stability analysis of the method.
dam to estimate extent of cracking
Unless the foundations of the dam are
and sliding stability. A set of inertial significantly stiffer than the concrete in the
forces should be calculated for the
dam (and can therefore be considered as
dam with the storage at minimum being rigid), it is essential that the
draw-off level to estimate the extent
foundations are included in any FEM. Usual
of cracking from the downstream practice is to give the foundation rock zero
toe/face. This cracking may limit the
density and appropriate stiffness parameters.
allowable cracking from the upstream
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 4.0 Assessment of Concrete Dams for Seismic Ground Motion

C4.4.5 Analysis in the Time Domain principal stresses). This will affect sliding
and overturning stability and should
Linear Elastic therefore be taken into consideration.
Constitutive models in some FEA code exist
Many commonly available finite element for doing this. Where horizontal cracking is
analysis (FEA) software packages include assumed, gap elements that allow frictional
the ability to determine the earthquake resistance (but only defined or zero tensile
response of a dam using either direct strength) should be used.
integration or superposition methods. The
former method is more accurate but requires In modelling the dam’s foundations,
greater computer resources. The method is appropriate consideration should be given to
less conservative than methods in the the geological/geotechnical model, especially
frequency domain. Linear elastic analysis in with regard to continuous defects along
the time domain should be undertaken prior which movement might occur.
to any non-linear analysis. As noted in
Section C4.4.4 it is important to include the It is essential that damping factors and
foundation rock in any FEM, with time- stiffness moduli are appropriately modelled
history analysis, the damping factor has to be such that the hysteretic behaviour of the dam

built into the analysis. This is different to a and foundations is properly captured.
response spectrum analysis where the input
data (the response spectrum) accounts for the C 4.5 Approach to Analyses and
damping. Acceptance Criteria
C4.5.1 During the Earthquake
A publication by the USBR (2006) on the
C4.5.1.1 Gravity Dams
state of practice for the non-linear analysis of FERC (2016) states that stresses induced
concrete dams provides up to date advice on during extreme earthquake loading will
dynamic material properties, loads and load induce stresses exceeding relevant material

combinations, damping, fluid elements, strengths. Consequentially, no acceptance

appropriate FEA methods, FEM philosophy criteria are given for the behaviour of a
and ways of dealing with cracking. gravity dam during the course of extreme
A publication by FEMA (2014) provides earthquake shaking. FERC emphasises the

advice on analyses that might be carried out need for post-earthquake static analysis.
as part of a risk assessment for a concrete ANCOLD (2013) and this guideline take the
dam. approach that pseudo-static methods are
useful and in general likely to be
Consideration should be given to the conservative, so if factors of safety are ≥
direction that a crack might propagate. For minimum required in Tables 6.1 and 6.2 of
cracking at the base of a dam, the crack is that Guideline the behaviour of the dam for
likely to propagate along the foundation that seismic ground motion is likely to be
interface. For cracking through the concrete satisfactory.
of the dam itself, a crack is likely to be
horizontal along a lift surface where the The USACE (2007) requires that during
tensile strength across the lift surface is less extreme earthquake shaking, acceptance
than that of the concrete. However, where based on linear elastic time-history analysis
there is no discernible difference in tensile depends upon the demand/capacity ratio
strength between the lift surface and the (DCR) for the dam and the cumulative time
concrete in general, the crack could excursions of non-linear behaviour. The
propagate in a downward direction as well as demand is defined as “the ratio of stress
in the downstream direction (following the demands to the static tensile strength of the

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 4.0 Assessment of Concrete Dams for Seismic Ground Motion

concrete”. The DCR is not to exceed 2 for a of the structure and leading to an
very limited cumulative non-linear duration uncontrolled loss of storage. While some
and reduces to 1 for a cumulative non-linear structural elements in the dam may fail,
duration of 0.75 second. The DCR method is overall failure of the dam (breaching) will
further described in USACE (2003b). not necessarily occur. It may be necessary to
remove structural elements that are likely to
C4.5.1.2 Arch Dams: fail in order to see how loads re-distribute in
The same principles apply to arch dams as the dam and whether they are likely to lead
apply to gravity dams. In addition, to further failure of structural elements.
consideration has to be given to whether or Consideration should also be given to the
not to specifically model the storage (as ductility of various critically loaded
against using empirical formulae to estimate structural elements. A number of slab and
hydrodynamic pressure distributions) in any buttress dams have relatively light
finite element, time-history analysis. This reinforcement in the struts for example and
will depend on the relative fundamental the cracking moment for the strut is greater
frequencies of the dam and the storage – than the moment capacity for the cracked,
refer Section If an arch dam has reinforced concrete section inferring a non-
massive gravity abutments (thrust blocks) ductile cross section..

finite element modelling has to be such that
stresses at the interfaces between the ends of C4.5.2 Post-Earthquake
the arch and the abutment units can be easily
integrated to produce thrust and shear forces C4.5.2.1 Gravity Dams:
acting on the abutment units. These forces No commentary.
are then used to carry out 3D stability
C4.5.2.2 Arch Dams:
analysis of the gravity abutment units. The
Due to the usually small distance from the
acceptance criteria for abutment blocks may
upstream heel to the downstream toe
be taken as for concrete gravity dams.
(compared to a gravity dam), the change in
The post-earthquake factors of safety should uplift force from a fully drained uplift

be as given in ANCOLD (2013) for extreme pressure distribution (or linear from the
load cases. upstream heel to the downstream toe) to a
cracked uplift pressure distribution (full
Additional details on the evaluation of arch headwater pressure within the cracked zone)
dams can be found in Jonker (2014). will be relatively small. However,

consideration should be given to the reduced

C4.5.1.3 Buttress Dams: sliding resistance at the foundation interface
Massive buttress dams (e.g. Burrinjuck Dam caused by cracking in the arch. The stability
in NSW) should use the acceptance criteria of the abutment sections of the arch or of any
given for gravity dams. The following gravity abutments should be analysed for
discussion relates to buttress dams that are post-earthquake loadings taking any cracking
more structural in nature (e.g.thin slab and occurring during the earthquake into account.
buttress dams).
Linear elastic FEA may be undertaken using
While the interaction between the face slabs very low elastic modulus elements along the
and the buttresses is generally statically crack(s).
determinate, the interaction between the
buttresses and the struts between the The abutments of an arch dam should be
buttresses is generally statically analysed considering any redistribution of
indeterminate. That is, there is a significant stresses within the arch due to cracking.
degree of redundancy in that part of the Consideration should be given to the 3D
structure. Consequently, a number of struts analysis of gravity abutment units and of
could potentially fail without causing failure
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 4.0 Assessment of Concrete Dams for Seismic Ground Motion

rock wedges in the abutment foundations that

have viable release surfaces.

C4.5.2.3 Buttress Dams:

Carry out a static post-earthquake stability
and strength analysis of the buttress dam
considering structural cracking that may have
occurred in face slabs and buttresses and
structural that might have occurred in the
struts between buttresses. Residual shear
strength parameters should be used where
appropriate. Consideration should be given
to each individual buttress foundation as
localised foundation properties could lead to
differential sliding deformations which in
turn, could lead to differential structural
effects within the various parts of the buttress
dam. Reference should be made to FERC

(1997, “Other Dams”) and to USACE
(2007). It should be noted that the uplift
pressure under the whole of a buttress
foundation will be determined by the
tailwater level.

C4.5.3 General
In the analysis, uplift pressures are likely
to have been represented in the finite
element model, by a set of self-
equilibrating pressures on a thin layer of

elements at the base of the dam. It

should be remembered that the stresses
obtained from the finite element analysis
in this case, will be total stresses. In

determining the effective stress

distribution through the dam, the
distribution of seepage (pore) pressures
through the dam will have to be
Post-earthquake analysis would usually
be required when the dynamic analysis
of the dam and its foundations during
earthquake shaking indicate:

· Cracking propagates such that post-

earthquake sliding stability will be
reduced due to reduced cohesion (if
cohesion is being relied upon).
· Cracking is such that shear strengths will
reduce to residual strengths post-
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 5.0 Assessment of Tailings Dams for Seismic Ground Motion

C5 ASSESSMENT OF TAILINGS DAMS Small raises of the tailings dam can usually
FOR SEISMIC GROUND MOTIONS be designed so there is sufficient rockfill or
other non strain-weakening or cracking
C5.1 Some General Principles material in the starter dams so that adequate
Figures 19.17 to 19.24 of Fell et al (2005, post-earthquake stability can be provided and
2015) show schematically the methods of deformations limited to tolerable amounts
construction. with a high degree of confidence. Essentially
these become centreline construction raises.
C5.2 The Effects of Earthquakes on
Tailings Dams C5.4 Design Seismic Loads and Analysis
See ANCOLD (2012) for more information. Method
This Guideline, which draws no distinction
C5.3 Defensive Design Principles for between conventional and tailings dams,
Tailings Dams recommends different design earthquake
These Guidelines caution against the use of loadings during operation to what is
upstream construction methods for High and recommended in Table 7 of ANCOLD
Extreme Consequence Category tailings (2012a) . The reasons for this are:
dams and require practitioners taking this

approach to use conservative assumptions (a) The OBE should be determined by
based on the following information: the Owner in consultation with the
Consultant in consideration of dam
(a) There have been a number of failures safety and business risks.
of such dams due to seismic loads. (b) The OBE in Table 7 of ANCOLD
(b) It is very difficult to ensure that (2012a) were not consistent with
quality control of placement of the usual practice in that it is seldom that
tailings during the operation of the an OBE with a frequency as low as 1
TSF (often many years) will be in 1000 AEP is adopted. OBE of 1 in
sufficient to give the required low 50 and 1 in 100 are so low a load as
likelihood of failure. One loose to make them not significant.

saturated layer in otherwise properly (c) The SEE (MDE) for TSF should be
densely placed tailings is sufficient to consistent with those for conventional
lead to liquefaction. dams. ANCOLD (2012a) has a lower
(c) There are large uncertainties in SEE for Low Consequence Category

predicting the liquefied strengths and dams (1 in 100).

deformations of liquefied tailings.
(d) It would be contradictory to say it is The post closure design earthquake loading
good practice to site conventional is shown in ANCOLD (2012a) as the MCE
dams to avoid liquefaction, and not for all Consequence Categories of TSF. This
recommend against upstream was in recognition of the long life after
construction for High and Extreme closure of a tailings dam compared to say
Consequence Category tailings dams. 100 years for a conventional dam.

This ultimately is a matter for State ANCOLD (2012a) also recommended the 1
Regulators to decide. ANCOLD counsel that in 1000 AEP ground motion for OBE to
if upstream or centreline construction is used better capture the potential for liquefaction of
for high and extreme consequence category tailings because they may liquefy at 1 in
dams persons expert in liquefaction of 1000 AEP ground motions and not at 1 in
tailings be involved; conservative lower 500 AEP ground motions.
bound strengths be adopted for design; and
These issues have been considered for this
assumptions made for design confirmed by
Guideline and this Guideline concludes:
testing during operation of the dam.

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 5.0 Assessment of Tailings Dams for Seismic Ground Motion

(a) In Australian conditions it is seldom should be re-evaluated at the time of closure

possible to identify the location and to allow for any developments in
activity of all faults which may understanding of seismic hazard and for any
contribute to the seismic hazard so changes in the consequences of failure, and
the design seismic loading should be that allowance be made for potential
expressed in probabilistic terms, not increases in the PAR post closure.
MCE. This is as is done in Table 2.1
of this Guideline. For Extreme C5.5 Seismic Deformation Analysis of
Consequence Category dams there is Tailings Dams not Subject to Liquefaction
also a requirement for MCE. Section and Figure 6 of ANCOLD
(b) The requirement of ANCOLD (2012a) incorrectly indicate that the
(2012a) for MCE post closure for Swaisgood (1998) and Pells and Fell (2003)
Low and Significant Consequence methods can be applied to TSF which
Category tailings dam could be experience liquefaction. Those methods are
considered onerous. It is inconsistent based on databases which exclude
with the requirement for conventional liquefaction cases. Swaisgood (1998) does
dams. include hydraulic fill dams but his plots do
(c) The argument that tailings dams are not include those which have liquefied.

required to operate post closure for an
C5.6 Assessment of the Effects of
indefinite period and therefore should
Liquefaction in Tailings Dams and
be designed for the largest credible
Their Foundations
earthquake has some merit but to
As described in ANCOLD (2012) Section
apply it to all tailings dams
6.1.5, tailings are usually stratified, so under-
contradicts normal risk based
drains may not result in complete drainage of
management because the
the tailings. Perched water tables with layers
consequences of failure for Low and
of saturated tailings should be assumed to be
Significant Consequence Category
present unless extensive investigations prove
tailings dams are lower than for High
and Significant Consequence

Category tailings dams. Also the C5.7 Methods for Upgrading Tailings
principles of tolerable risk criteria are Dams for Seismic Loads
that they are annual, not summed In mining operations there will often be a
over the life of a structure, so the life ready and relatively inexpensive supply of

of the structure is not a factor. waste rock to construct berms to improve

(d) In any case many conventional dams post-earthquake stability and reduce
are likely to operate for centuries. deformations to tolerable amounts.
(e) The OBE should not be used as a de
facto SEE to pick up liquefaction The geochemistry of the rock should be
issues. This Guideline specifically considered to ensure that there are no acid
says to consider liquefaction for the mine generation issues.
SEE of Significant and Low
Consequence Category dams and to
use the higher loadings if liquefaction
is assessed to be an issue.
(f) The principle that OBE is for Owners
to decide should stand.

Based on these factors it is concluded that

the post closure design seismic loading
should be as for the SEE, but that the SEE
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 6.0 Assessment of Appurtenant Structures for Seismic Ground

C6 ASSESSMENT OF APPURTENANT structurally interact with it. There can be

STRUCTURES FOR SEISMIC GROUND extreme variability in the amount of
reinforcement within existing towers not
designed for seismic conditions with steel to
C6.1 The Effects of Earthquakes on concrete areas often ranging from 0.2 percent
Appurtenant Structures to 2 percent, a variation factor of about 10. In
It is important that their functional and many existing towers, minimal or no
structural integrity of appurtenant structures confinement (transverse) steel is provided,
is retained in the event of a notable which is a very important aspect in the
earthquake. The importance of the role of the earthquake structural engineering. There are
appurtenant structures was particularly many cases where there will be a strong and
highlighted in the Wenchuan earthquake in a weak axis of the intake tower depending
China in 2008, as discussed in Wieland and upon the shape of the tower and the intake
Chen (2010), which included examples of: and outlet arrangement. It is recommended
that the tower is assessed in both directions.
· Outlet works control panels toppling All of this variability in the towers

leaving the outlet works inoperable; underscores the diversity of design
· Penstock failure and powerhouse encountered and the need to consider each
flooding due to rock falls; and intake tower individually, based on the actual
· Large scale rock falls and landslides configuration and applicable seismic loads.
leaving dams inaccessible for a
In the assessment it needs to be considered
number of months following the
that the frequency of the tower will change
earthquake event.
following the inclusion of added masses for
It is recommended that the reader also refers either or both the external and internal water.
to ICOLD which has a number of Guidelines
It is to be noted that in general, intake towers
relating to earthquake analysis and design of
have historically performed well during

appurtenant structures including ICOLD

earthquakes, with minimal reported instances
(2002) ‘Bulletin 123: Earthquake design and
of tower damage or failure.
evaluation of structures appurtenant to
dams’. C6.2.2 Defensive Design Principles

Regardless of the method of analysis No commentary.

selected, the final evaluation of seismic
safety of an appurtenant structure should be C6.2.3 Performance Requirements
based on engineering judgement and No commentary.
experience with similar structures, keeping in
mind that each structure and its immediate C6.2.4 Analysis Procedures
environment are unique and may not be Response Spectrum
duplicated elsewhere. This type of analysis assesses the maximum
response of the structure to earthquake
C6.2 Intake/Outlet Towers
excitation by combining the maximum
C6.2.1 General responses from individual modes and multi-
component inputs. This method uses an
The apparent simple geometry of intake added mass representation of hydrodynamic
towers is often complicated by the presence effects due to surrounding (outside) water
of various openings and appendages. There and contained (inside) water (in the case of
are also often changes in wall thickness at wet towers). In addition, it includes the
various elevations. Sometimes, the access effects of the tower/foundation interaction.
bridge that provides access to the tower will
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 6.0 Assessment of Appurtenant Structures for Seismic Ground

The determination of the hydrodynamic mass serve to demonstrate the general behaviour
of the water should be done using Goyal and of the seismic response, and can provide
Chopra (1989a,b,c,d,e). The Square-Root-of- some estimate of the expected inelastic
the-Sum-of-Squares (SRSS) method is behaviour or damage when the non-linearity
appropriate to combine the multidirectional is considered to be slight to moderate. Time-
loading components. Although, the Complete history analysis provides valuable time-
Quadratic Combination (CQC) method can dependant information that is not-available in
be used as an alternative to the SRSS the response spectrum analysis. Especially
method. A damping ratio of 5% is typically important is the number of excursions
adopted for concrete structures, for steel beyond displacement levels where the
structures 2% damping is typically adopted. structure might experience strength
The accuracy of the results depends on the degradation (strain softening).
number of vibration modes considered and
the methods of combination used for the Linear time-history analysis involves
modal and multi-component earthquake computation of the complete response history
responses. The response spectrum analysis of the structure to earthquakes. USACE

procedures are detailed in full in USACE (2007) provides details for the assessment
(2003a) and USACE (2007). and criteria using linear time-history
analysis. The basic procedure is to perform a
The response spectrum analysis is adequate linear time-history analysis with appropriate
for towers whose responses to earthquakes amount of damping to obtain bending
are within the linear elastic range. All towers moment DCR ratios for all finite elements.
should be designed to remain in the linear Initially a damping ratio of 5 percent is used
elastic range for the OBE. However, it will and then increased to 7 percent if DCR ratios
not generally be necessary for the tower to are approaching 2 and to 10 percent if they
behave elastically during the SEE. Therefore, exceed 2. After adjusting the damping the
in order to ensure that the tower can survive damage is compared against the criteria.

intense ground shaking due to the SEE with

limited damage, it should possess a ductility From the outcomes of the assessment the
capacity greater than ductility requirements damage is considered moderate and
imposed by the ground motion. This is acceptable if the following conditions are

expressed as a demand capacity ratio (DCR), met:

which is the ratio of the computed section · Bending moment DCR ratios computed
force demand to the section force capacity. on the basis of linear time-history
For an SEE the minimum DCR should be analysis remain less than 2;
taken as equal to 2 for flexure and 1 for · Cumulative duration of bending moment
shear. excursions above DCR ratios of 1 to 2
fall below the acceptance curve provided
Linear Time-History Analysis in USACE (2007); and
Linear time-history analysis involves · The extent of yielding along the height
computation of the complete response history of the tower (i.e. plastic hinge length for
of the structure to earthquakes. The DCR ratios of 1 to 2) is limited and falls
procedure for this analysis is similar to that below the acceptance curve provided in
described for the response spectrum USACE (2007).
procedure, except that earthquake demands If the DCR ratios exceed 2.0 or the
are in the form of acceleration time-histories, cumulative duration and the yield lengths
rather than response spectra and the results rise above the acceptance curves, the damage
are in terms of displacement and stress (or is considered to be severe and would need to
force) histories. The results of such analysis

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 6.0 Assessment of Appurtenant Structures for Seismic Ground

be assessed using non-linear analysis Stability Assessment - Rotational

A structure will tip about one edge of its base
Non-linear Time-History Analysis when earthquake plus static overturning
The non-linear analysis of towers is typically moment (Mo) exceed the structure restoring
complicated and time consuming. Until only moment capacity (Mr), or when the resultant
recently this type of analysis has been of all forces falls outside the base. Rocking
considered non-practical. However the responses to ground motions may include:
increase in capabilities of computers and the
· No tipping because Mo > Mr
further understanding of the performance of
these structures in earthquakes is leading to · Tipping or uplift because Mo > Mr, but no
more emphasis being placed on these types rocking due to insufficient ground motion
of analyses. For details on non-linear
analysis of intake tower refer to USACE · Rocking response (Mo > Mr) that will
(2007). eventually stop due to the energy loss
during impact
Stability Assessment - Sliding · Rocking response that leads to rotational

instability (considered unlikely)
Response Spectrum Method
The sliding stability for the response Tipping Potential Evaluation
spectrum method is expressed in terms of a When towers are subjected to large lateral
factor of safety. The response spectrum forces produced from an earthquake, the
analysis procedures described for the tower may start to tip and start rocking when
Strength Assessment above are adopted to the overturning moment becomes so large
determine the inertial forces on the structure. that the structure breaks contact with the
ground. The assessment of the tipping
Linear Time-History Analysis potential for the tower, either a rigid or
The results of the linear elastic time-history flexible tower, is provided in USACE

analysis can be used to compute time-history (2007). If it is determined that no tipping

or instantaneous sliding factor of safety occurs then the tower is considered stable. If
along a sliding plane. The procedures for this tipping occurs then the rocking block
assessment are provided in USACE (2007). analysis should be conducted.

Within this analysis, the stability is

maintained and sliding does not occur if the Rocking Block Analysis
factor of safety is greater than 1. However, a A tower would eventually overturn if the
factor of safety of less than 1 indicates a moment Mo > Mr is applied and sustained.
transient sliding, which if repeated numerous However, the overturning moments typically
times, could lead to excessive or permanent only occur for a fraction of a second in each
displacement. If this is the case then the cycle in a large earthquake, with intermediate
magnitude of the sliding displacement may opportunities to unload. Therefore, although
need to be evaluated using non-linear rocking occurs tower structures are unlikely
analysis. to become unstable. USACE (2007) provides
the procedures for assessment of the tower
Non-linear Time-History Analysis rocking.
In the non-linear time-history analysis an
assessment can be made of the total C6.3 Spillways
permanent sliding displacement of the tower. Gated spillway structures may be unstable
It then needs to be assessed whether the during seismic loading due to a combination
permanent displacements meet the criteria of large water loads concentrated near the top
for the SEE. of the structure and the relative lack of mass
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 6.0 Assessment of Appurtenant Structures for Seismic Ground

involved with this type of structure. C6.5.4 Analysis Procedures

Increasing the mass at the base of the Guidance is provided in FEMA (2005)
structure, anchoring the structure, or other regarding the seismic assessment of outlet
measures, may be necessary to stabilise the conduits. This document suggest that unless
structure during seismic loading. the conduit is founded upon deep soil layers,
C6.4 Spillway Gates seismic ground motions for the bedrock are
No commentary. typically assumed to act on the conduit. For a
conduit founded on soil the ground motion
C6.5 Outlet Works - Water Conduits, should allow for amplification or de-
Gates and Valves amplification effects. For a structure with a
fundamental frequency less than 33Hz, it is
C6.5.1 General suggested that a pseudo-static approach is
World-wide experience with the performance generally suitable. Where structures have a
of tunnels during earthquake has been very higher fundamental frequency, a response
good. Even soft ground tunnels have spectrum or time-history analysis may be
performed well as long as their design required.

incorporated some degree of articulation and
flexibility. This has been attributed to Information on the assessment and
decrease of the inertial effects versus performance of tunnels for earthquakes is
also provided in Wang (1993). Dynamic
kinematic ones in their seismic response and
also due to decreasing rate of seismic loads on the conduit are to include those
intensity in the ground depth. However, induced from the earthquake effect on the
particular attention is required at tunnel overlying dam.
portals. Since portals are surface features,
they are usually situated in weak rock, Earthquakes can cause substantial
experience surface wave reflections, and are hydrodynamic forces in penstocks and
exposed to blockage by earthquake-induced pressure tunnels depending upon the

rock falls and landslides. Most cases of foundation conditions and length of the
historic earthquake damage to tunnel systems penstock or tunnel. High dynamic pressures
have been at tunnel portals. Particular can occur in short sections of tunnel such as
consideration also needs to be given where those through the body of a dam. The reader

tunnels go from within rock to within an is referred to Weiland (2005) which provides
embankment. analysis procedures for the hydrodynamic
A comprehensive study of the effects of
earthquake and tunnels can be found in C6.6 Retaining Walls
Dowding and Rozen, 1978, and Owen and
Scholl, 1981. Further information on the C6.6.1 General
assessment procedures for tunnels for No commentary.
earthquakes is provided in Wang (1993).
C6.6.2 Defensive Design Principles
C6.5.2 Defensive Design Principles No commentary.
No commentary.
C6.6.3 Performance Requirements
C6.5.3 Performance Requirements No commentary.
No commentary.

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 6.0 Assessment of Appurtenant Structures for Seismic Ground

C6.6.4 Analysis Procedures using plane-strain 2-D soil elements whose

shear modulus and damping vary with level
Depending on the magnitude of wall
of shearing strains, and the non-linear
movements the backfill material is said to be
behaviour is approximated by the equivalent
in yielding, non-yielding, or intermediate
linear method.
state. Accordingly, the available methods of
design and analysis of the backfill soil C6.7 Parapet Walls
pressures also fall into similar categories. Parapet walls should be designed for Static
Further information on the analysis and dynamic loads to satisfy maximum
procedures is provided in USACE (2007). concrete and steel reinforcement stresses
which satisfy AS3600 (Concrete Structures).
Dynamic pressures of yielding backfill.
Yielding backfill condition means wall Detailing of the wall, and how the
movements due to earthquake ground embankment is zoned to support the wall is
motions are sufficient to fully mobilize shear critical to the performance under seismic
resistance along the backfill wedge creating ground motions.
limit state conditions. The dynamic earth

forces will then be proportional to the mass C6.8 Mechanical and Electrical
in the failure wedge times the ground Equipment
accelerations. When designing retaining No commentary.
walls with yielding backfill conditions for
earthquake ground motions, the Mononobe- C6.9 Access Roads and Bridges
Okabe (Mononobe and Matuo 1929; Okabe No commentary.
1924) approach and its several variations are
C6.10 Reservoir Rim Instability
often used. This is the most common
approach used in the assessment of retaining The assessment of landslides is a specialist
walls, refer to USACE (2007). area and suitably qualified engineering
geologists and geotechnical engineers should

Dynamic pressures of non-yielding be consulted.

backfill. For massive structures with soil
backfill, it is unlikely that movements The following references are relevant to the
sufficient to develop backfill yielding will assessment:
occur during an earthquake. In this situation

ICOLD Bulletin 124, Reservoir landslides:

the backfill soil is said to be non-yielding
investigation and Management-Guidelines
and is treated as an elastic material. If
idealized as a semi-infinite uniform soil and case histories. ICOLD (2002).
layer, the dynamic soil pressures and
Effect of earthquakes on landslides:
associated forces for a non-yielding backfill
can be estimated using a constant-parameter
· Keefer (1984)
single degree of freedom model or a more
elaborate multiple degree of freedom system. Methods for assessing whether landslides
The dynamic soil pressures for a more
will travel slowly or rapidly:
general non-yielding backfill soil can be
determined by finite-element procedures.
· Glastonbury and Fell (2008a, 2008b,
Intermediate case. The intermediate case in 2010)
which the backfill soil undergoes non-linear · Fell et al (2007)
deformations can be represented by finite · Glastonbury et al (2002)
element procedures using a soil-structure
interaction computer program. The
foundation and backfill soil are represented
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Commentary – 6.0 Assessment of Appurtenant Structures for Seismic Ground

Methods for predicting impulse waves

resulting from landslides into reservoirs:

· Heller et al (2009)
· Panizzo et al (2005)

It should be noted that only slides which

move rapidly after failure, and as they enter
the reservoir will result in waves. The
references above give guidance on whether a
slide will be rapid.


ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017

REFERENCES Soc. Am 97, 1350-1354.

Abaqus (2012) Allen, T.I., Cummins, P.R., Dhu, T. and
Schneider, J.F. (2007) Attenuation of
Abrahamson N.A. and Silva W.J. (2008). ground motion spectral amplitudes in
Summary of the Abrahamson and Silva south eastern Australia. Bull. Seism. Soc.
ground motion relations, Earthquake Am 97, 1279-1292.
Spectra 24, 67-98.
Allen, T. I., Gibson, G., Brown, A., and Cull,
Abrahamson N.A., G. Atkinson, D. J.P. (2004) Depth variation of seismic source
Boore, Y. Bozorgnia, K. Campbell, B. scaling relations: implications for earthquake
Chiou, I.M. Idriss, W. Silva and R. hazard in south eastern Australia,
Youngs (2008). Comparisons of the Tectonophysics, 390, 5-24.
NGA ground motion relations.
Earthquake Spectra 24, 45-66. ANCOLD (1998) “Guidelines for the Design
of Dams for Earthquake” Australian National
Abrahamson. N.A. and K. Shedlock, Committee on Large Dams.
1997. Overview of ground motion
attenuation relations. Seismological ANCOLD (2003). Guidelines on Risk

Research Letters, 68: 9-23. Assessment. Australian National Committee
on Large Dams, Hobart.
Abrahamson, N., (2008), Appendix C
Probabilistic Fault Rupture Hazard ANCOLD (2012), Guidelines on the
Analysis: San Francisco PUC, General Consequence Categories for Dams.
Seismic Requirements for the Design on Australian National Committee on Large
New Facilities and Upgrade of Existing Dams, Hobart.
Facilities, 1–7 p.
ANCOLD (2012a). Guidelines on Tailings
Abrahamson, N., W. Silva and R. Kamai Dams: Planning, Design, Construction,
(2014). Update of the AS08 Ground- Operation and Closure. Australian National

Motion Prediction Equations Based on Committee on Large Dams, Hobart.

the NGA-West2 Data Set. Earthquake
Spectra: August 2014, Vol. 30, No. 3, ANCOLD (2013). Guidelines on Design
pp. 1025-1055. Criteria for Concrete Gravity Dams.
Australian National Committee on Large

Allen, T. (2011), "The 2012 Australian Dams, Hobart.

Seismic Hazard Map – Catalogue and
Ground Motion Prediction Equations." ANCOLD (2016). Guidelines on Risk
Proceedings of the Australian Assessment. Australian National Committee
Earthquake Engineering Society on Large Dams, Hobart. In preparation.
Conference, 18-20 Nov, Novotel Barossa
Anderson, J. G., and Hough, S.E. (1984).
Valley Resort, South Australia.
A model for the shape of the Fourier
Allen, T.I. (2012). Stochastic ground-motion amplitude spectrum of acceleration at
prediction equations for south eastern high frequencies, Bull. Seism. Soc. Am.
Australian earthquakes using updated source 74, 1969-1993.
and attenuation parameters. Geoscience
Andrus, R.D., Hayati, H. and Mohanan, N.P.
Australia Record 2012/069.
(2009). Correcting liquefaction resistance for
Allen, T.I. and Atkinson, G.M. (2007). aged sands using measured to estimated
Comparison of earthquake source spectra velocity ratio. J. Geotech. Geoenviron.
and attenuation in eastern North America Engrg. ASCE. Vol.135, 735-744.
and south eastern Australia. Bull. Seism.
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017

Andrus, R.D. and Stokoe, K.H., II (2000). Beaty, M.H. and P.M. Byrne (2008).
‘Liquefaction resistance of soils from shear- “Liquefaction and Deformation Analysis
wave velocity’. J. Geotech. and Geoenvir. Using a Total Stress Approach.” ASCE
Engrg., ASCE, Vol. 126, No. 11, 1015– Journal of Geotechnical and
1025. Geoenvironmental Engineering, 134(8),
Arango, I., Lewis, M. R., and Kramer,
C._(2000)_. “Updated liquefaction potential Boore, D.M., J. P. Stewart, E. Seyhan,
analysis eliminates foundation retrofitting of and G, M. Atkinson (2014). NGA-
two critical structures.” Soil Dyn. Earthquake West2 Equations for Predicting
Eng., 20, 17–25. Response Spectral Accelerations for
Shallow Crustal Earthquakes.
Atkinson, G. M., and Boore, D.M. Earthquake Spectra: August 2014, Vol.
(2006). Earthquake ground-motion 30, No. 3, pp. 1057-1085.
predictions for eastern North America,
Bull. Seism. Soc. Am. 96, 2181-2205. Boulanger, R. W. (2010). A sand plasticity
model for earthquake engineering
AS1710 - Structural design actions, applications. Report No. UCD/CGM-10-01,

Standards Australia. Center for Geotechnical Modelling,
Department of Civil and Environmental
AS1170.4-2007: Structural design Engineering, University of California, Davis,
actions, Part 4: Earthquake actions in CA.
Australia. Standards Australia.
Boulanger, R.W., and Idriss., I.M. (2006)
AS3600 –Concrete Structure, Standards Liquefaction susceptibility criteria for silts
Australia. and clays, J. Geotechnical and
AS4100 –Steel Structures, Standards Geoenvironmental Engineering, ASCE, Vol
Australia. 132, No 11, 1413-1426.
Boulanger, R.W., and Idriss, I.M. (2007)

AS5100 – Bridge Design, Standards

Australia. Evaluation of cyclic softening in silts and
clays, Journal of Geotechnical and
Baker, J.W. (2010). An Introduction to Geoenvironmental Engineering, ASCE, Vol
the Conditional Mean Spectrum. 133, No6, 641-652.

Baker, J. W. (2011). Conditional Mean Boulanger, R.W. and Idriss, I.M (2012)
Spectrum: Tool for ground motion Probabilistic Standard Penetration Test –
selection. Journal of Structural based liquefaction –triggering procedure.
Engineering, 137(3), 322-331. Journal of Geotechnical and (ASCE) Geoenvironmental Engineering, ASCE,
ST.1943-541X.0000215 Vol138, No10, 1185-1195.

Baker, J.W. and Cornell, C.A. (2006). Boulanger, R. W. and Ziotopoulou, K.

Spectral shape, epsilon and record (2012). PM4Sand (Version 2): A sand
selection. Earthquake engineering and plasticity model for earthquake engineering
structural dynamics, 35, 1077-1095. applications. Center for Geotechnical
Modeling, Report No UCD.CGM-12-03.
Baziar, M.H. and Dobry, R. (1995). Residual Department of Civil and Environmental
strength and large scale deformation Engineering, University of California, Davis,
potential of loose silty sands. J. Geotechnical CA.
Engineering, ASCE, Vol. 121, No. 12, 896–
906. Bray, J.D. and T. Travasarou (2007).
“Simplified procedure for estimating
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017

earthquake induced deviatoric slope Geoenvironmental Eng., ASCE Vol 130, No

displacements”, Journal of Geotechnical and 12, 1314–340.
Geoenvironmental Engineering, ASCE, Vol.
133, No. 4, 381-392. Chiou, B. S.-J., and Youngs, R. R.
(2014). Update of the Chiou and Youngs
Bray , J.D and Sancio, R.B. (2006) NGA model for the average horizontal
Assessment of the liquefaction component of peak ground motion and
susceptibility of fine grained soils. J. response spectra, Earthquake Spectra 30,
Geotech. And Geoenvion. Eng. 132, 1117–1153.
Clark, D. and Van Dissen, R. (2006)
Brown, A. and Gibson, G. (2000). Characterisation of Australian intra-plate
Reassessment of earthquake hazard in active faulting with special mention of the
Australia, Proc 12th World Conf on Lake Edgar Fault, Tasmania. GNS Science
Earthquake Engineering, Auckland, Consultancy Report 2006/099
paper 751, 8 pages.
Clark, D., McPherson, A.and Van Dissen, R.
Brown, A. and Gibson, G. (2004). A multi- (2012). Long-term behaviour of Australian
tiered earthquake hazard model for Australia. Stable Continental Region (SCR) faults,

Tectonophysics 390, 25-43. Tectonophysics 566-567, 1-30.
Burbidge, D. (ed.) (2012). The 2012 Clark D., McPherson, A. and Collins C. D.
Australian Earthquake Hazard Map. N. (2011). Australia’s seismogenic
Geoscience Australia Record 2012/71. neotectonic record: a case for heterogeneous
intra-plate deformation. Geoscience
Burbidge, D and Leonard, M. (2011), Australia Record 2011/11.
The 2012 Australian Seismic Hazard
Map – Draft Maps. Australian Clough, R.W. and Ghanaat, Y. (1993)
Earthquake Engineering Society 2011 “Concrete Dams:Evaluation for Seismic
Conference, South Australia. Loading” Dam Engineering/ICOLD

International Workshop on Dam Safety

Campbell, K.W. and Bozorgnia, Y. Evaluation Vol. IV
(2014). NGA-West2 Campbell-
Bozorgnia Ground Motion Model for the Clough, R.W., Jenzien, J, (2003) “Dynamics
of Structures” 3rd Edition, pub. Computers &

Horizontal Components of PGA, PGV,

and 5%-Damped Elastic Pseudo- Structures Inc.
Acceleration Response Spectra for
Periods Ranging from 0.01 to 10 sec. Cox, B.R. and Griffiths, S.C. (2011).
Earthquake Spectra: August 2014, Vol. Practical recommendations for evaluation
30, No. 3, pp. 1087-1115. and mitigation of soil liquefaction in
Arkansas. Project number: MBTC 3017,
Campbell, K. W. (2009). Estimates of Mack-Blackwell Transportation Center,
shear-wave Q and Kappa for University of Arkansas.
unconsolidated and semiconsolidated
sediments in eastern North America, Crone, A.J., Machette, M.N., Bowman, J.R.,
Bull. Seism. Soc. Am. 99, 2365–2392. 1997. Episodic nature of earthquake activity
in stable continental regions revealed by
Cetin, K. O., Seed, R. B., Der Kiureghian, palaeoseismicity studies of Australian and
A., Tokimatsu, K., Harder,L. F., Kayen, R. North American Quaternary faults.
E., and Moss, R. E. S., (2004). Standard Australian Journal of Earth Sciences 44,
penetration test-based probabilistic and 203–214.
deterministic assessment of seismic soil
liquefaction potential, J. Geotechnical and
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017

Cubrinovski, M. and Ishihara, K. (2000a). Engineering Division, Vol. 104, No. 2,

Flow potential of sandy soils with different February 1978, pp. 175-191.
grain compositions. Soils and Foundations,
40(4): 103–119. Duron, Z.H., Ostrom,D.K Aagaard, B (1994)
“Measured Steady State, Ambient and
Cubrinovski, M. and Ishihara, K. (2000b). Transient Responses of a Small Arch Dam”,
Flow potential of sandy soils. Proc. GeoEng. Dam Engineering Vol IV, Issue 1
2000, Technomic Publishing, Lancaster,
549–550. Evernden, J. F., (1975) Seismic Intensities,
“size” of earthquakes and related parameters:
Cundall, P. (1993). FLAC – Fast lagrangian Seismological Society of America Bulletin,
analysis of continua, ITASCA. V. 65, p. 1287 – 1313.
Fell, R., Glastonbury, J. and Hunter, G.
Dafalias, Y. F. and Manzari, M. T. (2004). (2007) Rapid landslides: the importance of
Simple plasticity sand model accounting for understanding mechanisms and rupture
fabric change effects. Journal of Engineering surface mechanics. The eighth Glossop
Mechanics, ASCE, 130(6), 622-634.
Lecture. Quarterly Journal of Engineering
Geology and Hydrology, 40, 9-27.

Davidson, R.R, Fox, S. and Cooper, B.
(2003) Seismic upgrade of Yarrawonga
Fell, R., Macgregor, J.P., Stapledon, D.H.
Weir, ICOLD Congress.. ICOLD, Paris
and Bell. G. (2005) Geotechnical
Engineering of Dams. Balkema, Leiden.
DeJong, J.T., Ghafghazi, M., Sturm, A.,
Armstrong, R., Perez, A., and Davis, C. Fell, R., Macgregor, J.P., Stapledon, D.H.
(2014) “A New Instrumented Becker and Bell. G. and Foster, M (2015)
Penetration Test (iBPT) for Improved Geotechnical Engineering of Dams. Second
Characterization of Gravelly Deposits Within Edition. Taylor Francis, London.
and Underlying Dams, 2014, The Journal of
Dam Safety, Vol., 12, No. 2, pp. 9-19. FEMA (2000) “Pre-standard and

Commentary for the Seismic Rehabilitation

DeJong, J.T., Ghafghazi, M., Sturm, A.P.,
Wilson, D.W., den Dulk, J., Armstrong, R.J., of Buildings” Federal Emergency
Perez, A., and Davis, C.A. Management Agency, – 356, November
(2016a)“Instrumented Becker Penetration 2000.

Test, I: Equipment, Operation, and FEMA (2005) “Technical Manual: Conduits

Performance”, ASCE J. Geotechnical and through Embankment Dams” Federal
Geoenvironmental Engineering, in review. Emergency Management Agency.
Dise, K., Stevens, M.,G., and Von Thun, J.L. FEMA (2014) “Selecting Analytic Tools for
(1994), USBR Experiences Using DDC at 3 Concrete Dams Address Key Events Along
Dams for Mitigation of Liquefaction Risk, Potential Failure Mode Paths” FEMA P-
ASCE GSP No.45, 1-23. 1016, July 2014

Douglas, K.D. (2003). The shear strength of Fenves, G. and Chopra, A.K., “Simplified
rock masses. PhD Thesis. School of Civil Analysis for Earthquake Resistant Design of
and Environmental Engineering, The Concrete Gravity Dams”, College of
University of New South Wales, Sydney. Engineering, University of California,
Berkeley, Report No. UCB/EERC – 85/10,
Dowding, C.H. and Rozen, A. (1978) June 1986
“Damage to Rock Tunnels from Earthquake
FERC (2016) “Engineering Guidelines for
Shaking” Journal of the Geotechnical
the Evaluation of Hydropower Projects –
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017

Chapter 3 Gravity Dams”, Federal Energy Gravelly Soils for Liquefaction Potential
Regulatory Commission, March, 2016 Assessment at Dam Sites Using the iBPT”,
USSD 2014 Conference, San Francisco, CA,
FERC (1997) “Engineering Guidelines for 2014, pp. 285-297
the Evaluation of Hydropower Projects –
Chapter 10 Other Dams”, Federal Energy Ghafghazi, M., Thurairajah, A., DeJong, J.T.,
Regulatory Commission, October, 1997 Wilson, D.W., Armstrong, R. (2014b)
“Instrumented Becker Penetration Test for
Finn, W.D. (1988). Dynamic analysis, in Improved Characterization of Gravelly
Geotechnical Engineering Earthquake Deposits” GeoCongress 2014, GSP 234,
Engineering and Soil Dynamics II. Recent 2014, pp. 37-46.
Advances in Ground Motion Evaluation.
Geotech. Sp. Publ. No. 20, ASCE, 523–591. Ghafghazi M, DeJong JT, Sturm AP, Temple
CE. (2016)“Instrumented Becker Penetration
Finn, W.D.L. (1993). Seismic safety Test, II: iBPT- SPT correlation for
evaluation of embankment dams, in liquefaction assessment in gravelly soils”,
International workshop on dam safety ASCE J. Geotechnical and
evaluation, Grundewald, Switzerland, 26–27 Geoenvironmental Engineering, in review.

April, Vol. 4, 91–135.
Gibson, G. (1994) Earthqauke Hazard in
Finn, W.D.L. (2000). State-of-the art of Australi. Seminar, Acceptable Risks for
geotechnical earthquake engineering Extreme Events in the Planning and Design
practice. Soil Dynamics and Earthquake of Major Infrastructure. ANCOLD-Munro
Engineering, Vol 20, 1-15. Centre, 26-27 April 1994, Sydney.

Finn, W.D.L., Yogendrakumar, M., Yoshida, Glastonbury, J. and Fell, R. 2008a. A

N. and Yoshida, H. (1986). TARA-3: a decision analysis framework for the
program to compute the response of 2-D assessment of likely post failure velocity of
embankments and soil-structure interaction translational and compound natural rock
systems to seismic loadings. Department of slope landslides. Canadian Geotechnical

Civil Engineering, University of British Journal, Vol 45, No 3, 329-350.

Columbia, Canada.
Glastonbury, J.P. and Fell, R. 2008b. The
Finn, W.D.L., Kramer, S.L., O’Rourke, T.D., geotechnical characteristics of large slow,

and Youd, T.L. (2010). Ad Hoc Committee very slow and extremely slow landslides.
on soil liquefaction during earthquakes. Canadian Geotechnical Journal, Vol 45, No
Earthquake Engineering Research Institute 7, 984-1005.

Friesen, S. and Balakrishnan, A (2012). Glastonbury, J.P. and Fell, R. 2010

General approach used for the seismic Geotechnical characteristics of large rapid
remediation of Perris dam. Proceedings landslides, Canadian Geotechnical Journal,
USSD Conference, New Orleans, 1549-1564 Vol. 47, No 1, 116-132.

Foster, M and Fell, R. (2001) “Assessing Glastonbury, J., Fell, R. and Mostyn, G.
Embankment Dam Filters That Do Not (2002). Report on the Post Collapse
Satisfy Design Criteria”, Journal of Behaviour of Debris From Rock Slope
Geotechnical and Geoenvironmental Failures. UNICIV Report R-406, ISBN
Engineering, Vol . 127, No. 5, May, 2001. 85841 373 6, School of Civil and
Environmental Engineering, the University
Ghafghazi, M., DeJong, J.T., Sturm, A.P., of New South Wales.
Wilson, D.W., Davis, C., Perez, A., and
Armstrong, R. (2014a) “Characterization of
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017

Gomberg, J.S. (1992). Tectonic deformation Hall, L., Dimer, F and Somerville, P. (2007).
in the New Madrid seismic zone: inferences A Spatially Distributed Earthquake Source
from boundary element modelling. Model for Australia. Proceedings of the 2007
Seismological Research Letters, 63, 3, 107- Annual Meeting of the Australian
425. Earthquake Engineering Society.
Goyal, A. and Chopra, A.K. (1989a) Hansen, K.D. and Nuss, L.K. (2011)
“Earthquake Analysis and Response of “Lessons Learned from the Earthquake
Intake-Outlet Towers”, Earthquake Performance of Concrete Dams” Water
Engineering Research Centre, Report No. Power and Dam Construction, July 2011
UCB/EERC-89/04, University of California,
Berkeley. Hayati, H., and Andrus, R.D.,(2009)
Updated liquefaction resistance correction
Goyal, A. and Chopra, A.K. (1989b) factors for aged sands, Journal of
“Earthquake Analysis of Intake-Outlet Geotechnical and Geoenvironmental
Towers Including Tower-Water-Foundation- Engineering, American Society of Civil
Soil Interaction”, Earthquake Engineering Engineers, 135(11), p. 1683-1692.
and Structural Dynamics, 18(3), 325-344.

Heller, V., Hager, W.H. and Minor, H.E.
Goyal, A. and Chopra, A.K. (1989c) 2009. Landslide generated impulse waves in
“Earthquake Response Spectrum Analysis of reservoirs: Basics and computation. VAW
Intake-Outlet Towers”, Proc. ASCE, Journal Mitteilungen 211, 156pp.
of Engineering Mechanics, Vol. 115, No. 7.
Hoek, E. and Brown, E.T. (1997). Practical
Goyal, A. and Chopra, A.K. (1989d) estimates of rock mass strength. International
“Hydrodynamic and Foundation Interaction Journal of Rock Mechanics and Mining
Effects in Dynamics of Intake Towers: Sciences, 34(8), 1165–1186.
Frequency Response Functions”, Proc.
ASCE, Journal of Structural Engineering, Hoek, E. Marinos, P. and Benissi, M. (1998)
Vol. 115, No. 6. Applicability of the geological strength index

(GSI) classification for very weak and

Goyal, A. and Chopra, A.K. (1989e) sheared rock masses. The case of the Athens
“Simplified evaluation of Added Schist Formation. Bulletin of Engineering
Hydrodynamic Mass for Intake Towers”, Geology and the Environment, 57, 151–160.

Proc. ASCE, Journal of Engineering

Mechanics, Vol. 115, No. 7. Hossain, M.D. (2010) Correcting
Gregor, N., Abrahamson, N.A., liquefaction resistance of partially saturated
Atkinson, G.M., Boore, D.M., soil using compressional wave velocity. M.
Bozorgnia, Y., Campbell, K.W., Chiou, Sc. Thesis, Graduate School of Clemson
B.S-J. I. M. Idriss, R. K.,Seyhan, University.
E.,Silva, W., Stewart. J.P. and Youngs,
R. (2013) Comparison of NGA-West2 Hough, S.E., Armbruster, J.G., Seeber, L.,
GMPEs. Earthquake Spectra 30:3, 1179- and J.F. Hough (2000). On the Modified
Mercalli intensities and magnitudes of the
1811-1812 New Madrid earthquakes, J.
Gulerce, Z. and Abrahamson, N. (2011). Geophys. Res., 105-B10,23839-23864.
Site-Specific Design Spectra for Vertical
Ground Motion. Earthquake Spectra: Hudson, M.B., Martin, B.,Beikae, M.and
November 2011, Vol. 27, No. 4, pp. 1023- Idriss, I.M. (1994). “QUAD4M, a Computer
1047. Program to Evaluate the Seismic Response
of Soil Structures Using Finite Element
Procedures and Incorporating a Compliant
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017

Base”, NISEE, University of California, ICOLD 2002. Reservoir landslides –

Berkeley, California. guidelines for investigation and
management, International Commission on
Hunter, G. and Fell, R. (2003). Travel Large Dams, Paris. Bulletin 124.
distance for rapid landslides in constructed
and natural slopes. Canadian Geotechnical ICOLD (2012) Selecting seismic parameters
Journal, Vol. 40, No. 6, 1123–1141. for large dams - Guidelines (revision of
Bulletin 72).Bulletin 148, International
Hynes-Griffin, M.E. and A.G. Franklin Commission on Large Dams, Paris.
(1984). “Rationalizing the Seismic
Coefficient Method”, Miscellaneous Paper ICOLD (2013) Guidelines for Use of
GL-84-13, US Army Engineer Waterways Numerical Models in Dam Engineering.
Experiment Station, Vicksburg, MS. Bulletin 155, International Commission on
Large Dams, Paris.
Hynes, M.E., and Olsen, R.S. (1999).
Influence of confining stress on liquefaction ICOLD (2016a) Selecting Seismic
resistance. Proc., Int. Workshop on Phys. Parameters for Large Dams - Guidelines,
And Mech. Of Soil Liquefaction, Balkema, Bulletin 148, International Committee on
Large Dams, Paris.

Rotterdam, The Netherlands, 145–152.

Iai, S., and Koizumi, K., (1986). Estimation ICOLD (2016b) Inspection of Dams
of earthquake induced excess pore water Following Earthquakes, Bulletin 62A,
pressure for gravel drains, in Proceedings, 7th International Committee on Large Dams,
Japan Earthquake Engineering Symposium, Paris.
679–84. Idriss 2014, reference to be added.
ICOLD (1983). Seismicity of Dams and Idriss, I.M. and Boulanger, R.W, (2008) Soil
Design. Bulletin 46, International liquefaction during earthquakes. Monograph
Commission on Large Dams. Paris. MNO-12, Earthquake Engineering Research

ICOLD (1986). Earthquake analysis Institute, Oakland, Ca.

procedures for dams – State of the art. Idriss, I.M. and Boulanger, R.W, (2010) SPT
Bulletin 52, International Commission on –Based Liquefaction Triggering Procedures.
Large Dams. Paris. Report No UCD/CGM-10/02, Center for

ICOLD (1998) Neotectonics and Dams Geotechnical Modelling. University of

Recommendations and Case Histories. California at Davis.
Bulletin 112, International Commission on Idriss, I.M., Lysmer, J., Hwang, R. and Seed,
Large Dams. Paris. H.B. (1973). QUAD-4: A computer program
ICOLD (1999a). Design Features of dams to for evaluating the seismic response of soil
resist seismic ground motion. Bulletin 120, structures by variable damping finite element
International Commission on Large Dams, procedures. Report No. EERC73-16,
Paris. University of California, Berkeley.

ICOLD (1999b) Seismic Observations of Idriss and Seed (2010,2012) , reference to be

Dams. Bulletin 113, International added.
Commission on Large Dams. Paris.
Ishihara, K. (1993). Liquefaction and flow
ICOLD (2002) Seismic Design and failures during earthquakes. Geotechnique,
Evaluation of Structures Appurtenant to Vol. 33, No. 3, 351–415.
Dams, Bulletin 123, International Committee
on Large Dams.
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017

Itasca (2011) UBCSAND Constitutive assessment of liquefaction potential. Soil

Model on Itasca UDM Web Site by M. H. Mechanics Series No.38, Department of Civil
Beaty and P. M. Byrne. Engineering, University of British Columbia,
Vancouver, Canada.
Johnson, A.C., and Shedlock, K.M. (1992).
Overview of Research in the New Madrid Leon. E., Gassman. S.L. and Talwani. P.
seismic zine: inferences from boundary (2006). Accounting for soil aging when
element modelling. Seismology Research assessing liquefaction potential. J. of
Letters, 63, 3 193-208. Geotechnical and Geoenvironmental
Engineering, Vol. 132, No. 3, 363-377.
Jonker, M and Espander, R. (2014).
Evaluation of Existing Arch Dam Design Leonard M. (2010a). Earthquake Fault
Criteria in Lieu of ANCOLD Guidelines. Scaling: Relating Rupture Length, Width,
Average Displacement, and Moment
Kawai, T. (1985). Summary report on the Release. Bulletin of the Seismological
development of the computer program Society of America 100, 1971-1988.
DIANA – Dynamic interaction approach and
non-linear analysis. Science University of Leonard, M., T.I. Allen, D.R. Burbidge, D.J.

Tokyo. Clark, A. A. McPherson and C.D.N. Collins
(2012). The 2012 Australian Earthquake
Kayen, R., Moss, R. E. S., Thompson, E., M, Hazard Map. Geoscience Australia Record
Seed, R. B., Cetin, K. O., Der Kiureghian, 2012/Xx
A., Tanaka Y. and Tokimatsu, K. (2012)
Shear-Wave Velocity–Based Probabilistic Liang, J. Z., Hao,H., Gaull, B.A.and
and Deterministic Assessment of Seismic Sinadinovski, C. (2008). Estimation of
Soil Liquefaction Potential. J. of strong ground motions in southwest
Geotechnical and Geoenvironmental Western Australia with a combined
Engineering, Vol. 139, No. 3, 407–419 Green's function and stochastic
approach, J. Earthq. Eng. 12, 382–405.
Keefer, D.K. (1984) Landslides caused by

earthquakes. Geological Society of America Lysmer, J., Udaka, T., Tsai, C.F. and Seed,
Bulletin, Vol 95, 406-421. H.B. (1975). FLUSH – A computer program
for approximate 3-D analysis of soil-
Kempton, J.J. and J.P. Stewart [2006].
structure interaction problems. Report No.

Prediction Equations for Significant Duration

EERC75-30, University of California,
of Earthquake Ground Motions Considering
Site and Near-Source Effects,” Earthquake
Spectra, 22(4), pp. 985–1013. Luehring R, N Snorteland, L Mejia, and M.
Khalili, N., Fell, R. and Tai, K.S. (1996). A Stevens (2001): Liquefaction mitigation of a
simplified method for estimating failure silty dam foundation using vibro-stone
induced deformation. Proc. Seventh Int. columns and drainage wicks: a case history
Symp. On Landslides. Editor Senesett, K., at salmon lake dam. Bureau of Reclamation
Balkema, Rotterdam, 1263-1268. Water Operation and Maintenance Bulletin,
No. 198, 1-15.
Yamamuro, J.A. and Lade, P.V. (1997).
Static liquefaction of very loose sands. Makdisi, F.I. and Seed, H.B. (1978). A
Canadian Geotech. J. 34, 905-917. simplified procedure for estimating dam and
earthquake induced deformations. JASCE
Lee, M.K. and Finn, W.D.L. (1978). Geotechnical Eng., Vol. 1, 105, No. GT7,
DESRA-2, Dynamic effective stress 849–867.
response: Analysis of soil deposits with
energy transmitting boundary including
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017

Marcuson, W.F., Hynes, M.E. and Franklin, Earthquakes, Proceedings, World

A.G. (1990). Evaluation and use of residual Engineering Congress, 9.
strength in seismic safety analysis of
embankments. Earthquake Spectra, Vol. 6, Moriwaki, Y., Beikae, M. and Idriss, I.M.
No. 3, 529–572. (1988). Non-linear seismic analysis of the
Upper San Fernando Dam under the 1971
Marcuson, W.F., Hynes, M.E. and Franklin, San Fernando earthquake. Proc. 9th World
A.G. (1992). Seismic stability and permanent Conference on Earthquake Engng., Tokyo
deformation analysis: the last twenty five and Kyoto, Japan, Vol. III, 237–241.
years. Proc. ASCE Specialty Conference on
Stability and Performance of Slopes and Moss, R. E. S., Seed, R. B., Kayen, R. E.,
Embankments – II, Geotech. Special Publ. Stewart, J. P., Der Kiureghian, A., and Cetin,
No. 31 (eds R.B. Seed and R.W. Boulanger, K. O., (2006). CPT-based probabilistic and
ASCE, New York, Vol. 1, 552–592. deterministic assessment of in situ seismic
soil liquefaction potential, J. Geotechnical
Martin, G.R., Finn, W.D.L. and Seed, H.B. and Geoenvironmental Eng., ASCE 132(8),
(1975). Fundamentals of liquefaction under 1032–051.
cyclic loading. Journal of the Geotechnical

Engineering Division, ASCE, Vol. 101, 423– Motazedian, D., and Atkinson, G.M.
438. (2005). Stochastic finite-fault modelling
based on a dynamic corner frequency,
McCue, K. and M. Michael-Leiba (1993). Bull. Seism. Soc. Am. 95, 995–1010.
Australia’s deepest known earthquake.
Seismological Research Letters 64 (3-4). Newmark, N.M. (1965). Effects of
earthquakes on dams and embankments. 5th
Mcpherson, A.A. and Hall, L.S. (2007) Rankine Lecture, Geotechnique, Vol. 15, No.
Development of the Australian Regolith 2, 139–160.
Site Classification Map. Geoscience
Australia Record 20067/07 Nakazawa, H., Ishihara, K, Tsukamoto, Y,
and Kamata, T. (2004) Case studies on

Mejia, L. (2005). Seismic Retrofit of evaluation of liquefaction resistance of

Embankment Dams to Mitigate imperfectly saturated soil deposits. Proc. Int.
Liquefaction: Seleted Case Histories. Conf. on Cyclic Behavior of Soils and
Earthquake Engineering and Soil Liquefaction Phenomena. Bochum,

Dynamics: 1-14. Germany, Taylor and Francis Group, 295-

304.NSFW (1998). Shear strength of
McKay, M. and Lopez, F. (2015) liquefied soils. Proceedings of National
“Seismic Analysis of a Concrete Dam Science Foundation Workshop, Editors T.D.
with a Complex Configuration: Stark, S. Molson, S.L. Kramer and T.L.
Nonlinear, Time-History of Mt Bold Youd. National Science Foundation,
Dam” ANCOLD Proceedings 2015 Earthquake Hazard Mitigation Program
Mitchell, J. K., (2008). Mitigation of Okabe, S. (1924) General theory on earth
liquefaction potential of silty sands, in From pressure and seismic stability of retaining
Research to Practice in Geotechnical wall and Dam, Journal of the Japanese
Engineering, Laier, J.E., Crapps, D.K., Society of Civil Engineering, Vol. 10, No. 6
Hussein, M.H., (eds.), Geotechnical Special pp1277-1323, 1924.
Publication 180, ASCE, Reston, VA.
Olsen, S.M. and Stark, T.D. (2002).
Mononobe, N., and Matsuo, H. (1929), On Liquefied strength ratio from liquefaction
the Determination of Earth Pressures During flow failure case histories. Canadian

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017

Geotechnical Journal, Vol. 39, No. 3, 629– Prevost, J.H. (1998). “DYNAFLOW, Version
647. 98, Release 02.A”, Princeton University,
Department of Civil Engineering & Op. Res.,
Onoue, A., 1988. Diagrams considering well Princeton, New Jersey.
resistance for designing spacing ratio of
gravel drains, Soils and Foundations 28(3), Quigley , M.C., Cupper, M.L., and
160–68.Panizzo, A., De Girolamo, P, and Sandiford, M. (2006) Quaternary faults of
Petaccia, A. 2005. Forecasting impulse south eastern Australia: palaeoseismicity,
waves generated by subaerial landslides. slip rates and origin. Australian Journal of
Journal of Geophysical research 110 Earth Sciences 53, 285-301.
C12025: 1-23.
Rafael, J.M. (1984) “Tensile Strength of
Pells, S. and Fell, R. (2002). Damage and Concrete” ACI Journal March-April 1984
cracking of embankment dams by
earthquakes, and the implications for internal Rezaeian, S., Bozorgnia, Y., Idriss, I. M.,
erosion and piping. UNICIV Report R 406, Campbell, K., Abrahamson, N., and Silva,
The University of New South Wales, ISBN W., 2014a. Damping scaling factors for
85841 3752. elastic response spectra for shallow crustal
earthquakes in active tectonic regions:

Pells, S. and Fell, R. (2003). Damage and “Average” horizontal component,
cracking of embankment dams by earthquake Earthquake Spectra 30, 939–963.
and the implications for internal erosion and
piping. Proceedings 21st International Rezaeian, S., Bozorgnia, Y., Idriss, I. M.,
Congress on Large Dams, Montreal. ICOLD, Abrahamson, N., Campbell, K., and Silva,
Paris Q83–R17, International Commission W., 2014b. Damping scaling factors for
on Large Dams, Paris. vertical elastic response spectra for shallow
crustal earthquakes in active tectonic regions.
Perlea, V.G. and Beaty, M.H. (2010) Corps Earthquake Spectra 30, 1335–1358.
of Engineers practice in the evaluation of
Richter C.F., 1935, An instrumental

seismic deformation of embankment dams.

In Fifth International Conference on Recent earthquake magnitude scale, Bull.
Advances in Geotechnical Earthquake Seismological Society Am., 25: 1-32.
Engineering and Soil Dynamics, San Diego. Robertson, P.K. (2010). Evaluation of flow

Pestana, J. M., Hunt, C. E. and Goughnour, liquefaction and liquefied strength using the
R. R., (1997). FEQDrain: a Finite Element Cone Penetration Test. ASCE J. Geotech.
Computer Program for the Analysis of the Geoenviron. Eng. Vol. 136, No 6, 842-853
Earthquake Generation and Dissipation of Robertson, P.K. and Fear, C.E. (1995).
Pore Water Pressure in Layered Sand Liquefaction of sands and its evaluation.
Deposits with Vertical Drains, Report No. Proc., 1st Int. Conf. On Earthquake Geotech.
EERC 97-17, Earthquake Engineering Res. Engrg.
Ctr., University of California at Berkeley
Robertson, P.K. and Wride, C.E. (1997).
PLAXIS (2012) Delft, The Netherlands E- Cyclic liquefaction and its evaluation based
mail: on the SPT and CPT. Proc. NCEER
Workshop on Evaluation of Liquefaction
Prevost, J.H. (1981). DYNAFLOW: A non- Resistance of Soils. Editors T.L. Youd and
linear transient finite element analysis I.M. Idriss, NCEER Technical Report 97-
program. Princeton University, Department 0022, 41–88.
of Civil Engineering, Princeton, N.J.
Robertson, P.K. and Wride, C.E. (1998).
Evaluating cyclic liquefaction potential using
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017

the cone penetration test. Can. Geotech. J., Seed, H.B. and De Alba, P. (1986). Use of
Ottawa, 35(3), 442–459. the SPT and CPT tests for evaluating the
liquefaction resistance of sands. In-Situ 86,
Robertson, P.K. 2010, Evaluation of flow Conference on use of in-situ tests in
liquefaction and liquefied strength using geotechnical engineering. Clemence, S.P.
the Cone Penetration Test. ASCE J. (ed.). ASCE Geotechnical Special
Geotech. Geoenviron. Eng. Vol. 136, No. Publication, No. 6, 281–302.
6, 842-853.
Seed, H.B. and Idriss, I.M. (1982). Ground
Roth, W.H. (1985). Evaluation of earthquake Motions and Soil Liquefaction During
induced deformations of Pleasant Valley Earthquakes, Monograph Series, Earthquake
Dam. Report for the City of Los Angeles. Engineering Research Institute. Berkeley,
Dames and Moore, Los Angeles. California.
Sadigh, K., Chang, C.-Y., Egan, J.A., Seed, H.B. and Idriss, I.M. (1971).
Makdisi, F. and Young, R.R. 1997. Simplified procedure for evaluating soils
Attenuation relationships for shallow crustal liquefaction potential. J. Geotech. Engrg.
earthquakes based on California strong Div., ASCE, 97(9), 1249–1273.

motion data. Seismological Research Letters
68, 180-189. Seed, H. B. and Booker, J. R., (1977).
Stabilization of potentially liquefiable sand
Sandiford, M., Wallace,N. and Coblentz, N. deposits using gravel drains, J. Geotechnical
(2004). Origin of the in situ stress field in Eng. Div., ASCE 103 (GT7), 757–68.
southeastern Australia. Basin Research 16,
325-338. Seed, H.B., Lee, K.L., Idriss, I.M. and
Makdisi, F.I. (1973). Analysis of slides in the
Schnabel, P.L., Lysmer, J. and Seed, H.B. San Fernando Dams during the earthquake of
(1972). SHAKE: A computer program for February 1971. Report No. EERC/73-2,
earthquake response analysis of horizontally University of California, Berkeley.
layered sites. Report No. EERC72-12,

University of California, Berkeley. Seed, H.B., Tokimatsu, K., Harder, L.F. and
Chung, R.M. (1985). The influence of SPT
Schwartz, D.P. and Coppersmith, K.J. procedures in soil liquefaction resistance
(1984). Fault behavior and characteristic evaluations. J. Geotech. Engrg., ASCE,

earthquakes: Examples from the Wasatch 111(12), 1425–1445.

and San Andreas Fault Zones. Journal of
Geophysical Research, 89, Pages 5681– Seed, R.B. (2010) Technical review and
5698 comments: 2008 EERI monograph titled Soil
liquefaction during earthquakes,
Scott, G.A. and Van Thun, J.L. (1993)
Geotechnical Report No, UCB/GT-2010/01,
“Interim Guidelines-Geotechnical
University of California, Berkeley.
Studies for Concrete Dams” USBR
January 1993 Seed, R.B., Cetin, K.O., Moss, R.E.S.,
Kammerer, A., Wu, J., Pestana, J., Riemer,
Seed, H.B. (1979). Soil liquefaction and
M., Sancio, R.B., Bray, J.D., Kayen, R., and
cyclic mobility evaluation for level ground
Faris, A. (2003) Recent advances in soil
during earthquakes. J. Geotech. Engrg. Div.,
liquefaction engineering: A unified and
ASCE, 105(2), 201–255.
consistent framework, Earthquake
Seed, H.B. (1983). Earthquake-resistant Engineering Research Center Report No
design of earth dams. Proc., Symp. Seismic EERC 2003-06.
Design. Of Earth Dams and Caverns, ASCE,
New York, 41–64.
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017

Seed, R. B, and Harder, L. F.(1990). SPT- Earthquake Engineering Building an

based analysis of cyclic pore pressure Earthquake-Resilient Society, 14-16
generation and undrained residual strength, April, 2011, Auckland, New Zealand
in Proceedings, Seed Memorial Symposium,
J. M. Duncan, ed., BiTech Publishers, Stark, T.D., Beaty, M.H., Byrne, P.M.
Vancouver, British Columbia, 351–76. Castro, G., Walberg, F.C., G. Perlea, V.G.,
Axtell, P.J., Dillon, J.C., Empson, W.B. and
Sherard, J.L. (1967). Earthquake David L. Mathews D.L.(2012). Seismic
considerations in earth dam design. Journal deformation analysis of Tuttle Creek Dam.
of the Soil Mechanics and Foundations Can. Geotech. J. 49: 323–343 (2012).
Division, American Society of Civil
Engineers, Vol. 93, No. SM4, 377–401. Swaisgood, J.R. (1998). Seismically-induced
deformation of embankment dams. 6th US
Somerville, P.G., and G. Gibson (2008). National Conference on Earthquake
Seismic hazard at Hinze Dam. Engineering, Seattle, Washington.
Proceedings of the ANCOLD 2008
Conference, Gold Coast, Queensland, Swanson (1992). A computer program
November 12-15. designed by Swanson Analysis Systems, Inc.

Somerville, P.G., R.W. Graves, N.F. Collins, Sy A. and Campanella R.G. (1994).
S.G. Song, S. Ni and P. Cummins (2009). Becker and standard penetration tests
Source and ground motion models of (BPTSPT) correlations with
Australian earthquakes. Proceedings of the consideration of casing friction.
2009 Annual Conference of the Australian Canadian Geotechnical Journal, 31(3):
Earthquake Engineering Society, Newcastle, 343-356.
December 11-13.
Toose, J., Hunter, G. and Watkins, G
Somerville, P.G. and H.K. Thio (2011). (2007). Design and Construction of the
Epistemic Uncertainty in Seismic Hazard Lance Creek Reservoir Seismic Upgrade,
Analysis. Proceedings of the 2011 NZSOLD/ANCOLD Conference on

ANCOLD Conference, Melbourne, October Dams.

Travasarou, T., Bray, J.D., and Abrahamson,
Somerville, P.G. (2009). Probabilistic N.A. [2003]. “Empirical attenuation

seismic hazard analysis. Proceedings of the relationship for Arias Intensity,” Earthquake
Australian Geomechanics Society 2009 Engineering and Structural Dynamics; 32,
Seminar, Seismology and Earthquake 1133–1155.
Engineering, Adelaide, September 24.
Tsukamoto, Y. Ishihara, K., Nakazawa, H.,
Somerville, P., P. Quijada, H.Kie Thio, M. Kamada, K., and Huang, Y. (2002)
Sandiford, and M. Quigley (2008). Resistance of partially saturated sand to
Contribution of Identified Active Faults to liquefaction with reference to longitudinal
Near Fault Seismic Hazard in the Flinders and shear wave velocities. Soils and
Ranges. Proceedings of the 2008 Annual Foundations, 42,(6), 93-104.
Conference of the Australian Earthquake
Engineering Society, Adelaide, November USACE (1999) EM 1110-2-6050 Response
11-13. Spectra and Seismic Analysis of Hydraulic
Structures United States Army Corps of
Somerville, P.G. and Thio, H.K. (2011). Engineers.
Development of ground motion time-
USACE (2003a) EM 1110-2-2400 Design of
histories for seismic design. Proceedings
Outlet Works. United States Army Corps of
of the Ninth Pacific Conference on
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017

USACE (2003b) EM 1110-2-6051 Time- Watson-Lamprey, J. and Abrahamson, N.

History Dynamic Analysis of Hydraulic [2006]. “Selection of ground motion time
Structures. United States Army Corps of series and limits on scaling.” Soil Dynamics
Engineers. and Earthquake Engineering, Vol. 26, 477–
USACE (2005) Determination of Standard
Response Spectra and Effective Peak Ground Wesnousky, S.G., C. Scholz,
Accelerations for Seismic Design and C.H.,Shimazaki, N. and
Evaluation USACE, ERDC/CHL CHETN- Matsuda,T.(1983). Earthquake frequency
VI-41, December distribution and the mechanics of
faulting. Journal of Geophysical
USACE (2007) EM 1110-2-6053 Research 88, 9331-9340.
Earthquake Design and Evaluation of
Concrete Hydraulic Structures. United States Westergaard, H.M. (1933) Water Pressure
Army Corps of Engineers. on Dams during Earthquakes, ASCE
Transactions, Vol 98, pp 418-472.
USBR (2011) “Seismic Induced Loads on
Spillway Gates: Phase 1 Literature Review” Wieland, M. (2005). Hydrodynamic
USBR Report DSO-11-06, September, 2011. pressures, Int. Journal Water Power and Dam

Construction, October, pp 18-24.
USBR (2006) State-of-Practice for the Non-
linear Analysis of Concrete Dams at the Wieland, M. and Chen, H. [2010] Lessons
Bureau of Reclamation US Bureau of learnt from observed damage of dams and
Reclamation, Denver, Co. power plants caused by the May 12, 2008
Wenchuan Earthquake, The Dams
USBR (1989). Design Standards. Newsletter, February 2010, No. 9.
Embankment dams, No. 13, Chapter 13.
Seismic design and analysis. US Bureau of Yang, J., Savidis, S., and Roemer, M. (2004)
Reclamation, Denver, Co. Evaluating liquefaction potential of partially
saturated sand. J.Geotech and Geoenvir.

USNRC (1985). Liquefaction of soils during Eng. 130 (9), 975-979

earthquakes. National Academy Press,
Washington DC. Youd, T.L. (2011) A look inside the debate
over EERI Monograph MNO 12, Brigham

Van Houtte, C., Drouet, S., and Cotton, Young University. Available from:
F. (2011). Analysis of the origins of k
(kappa) to compute hard rock to rock 2011.pdf
adjustment factors for GMPEs, Bull.
Seism. Soc. Am. 101, 2926–2941. Youd, T.L., and Perkins, D.M. (1978)
Mapping of liquefaction-induced ground
Wang, J.N., (1993) Seismic Design of failure potential, Journal of Geotechnical
Tunnels – A simple State-of-the-Art Design Engineering, ASCE Vol104, No4, 433-446.
Approach. Parsons Brinkerhoff Quade and
Douglas Monograph 7. Youd, T.L., Idriss, I.M., Andrus, R.D.,
Arango, I., Castro, G., Christian J.T., Dobry,
Wang, Z.L. and Makdisi, F.I. (1999). R., Finn, W.D.L., Harder, L.F., Hynes, M.E.,
Implementing a bounding surface hypo Ishihara, K., Koester, J.P., Liao, S.S.C.,
plasticity model for sand into the FLAC Marcuson, W.F., Martin, G.R., Mitchell,
program, FLAC and Numerical Modelling in J.K., Moriwaki, Y., Power, MS., Robertson,
Geomechanics, A.A. Balkema, Netherlands, P.K., Seed, R.B. and Stokoe, K.H. (2001).
483-490. Liquefaction resistance of soils: summary
report from the 1996 NCEER and 1998
NCEER/NSF Workshops on evaluation of
ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017

liquefaction resistance of soils. J. Zienkiewicz, O.C., Valliappan, S. and King,

Geotechnical and Geoenvironmental I.P. (1986). Stress analysis of rock as a non-
Engineering, ASCE, Vol. 127, No. 10, 817– tension material. Geotechnique, 18, 55–56.
Ziotopoulou, K, and Boulanger, R.W.
Zangar, C.N. (1952) “Hydrodynamic (2012). Constitutive modeling of duration
Pressures on Dams Due To Horizontal and overburden effects in liquefaction
Earthquake Effects” Engineering evaluations. Proc. Second Int. Conf. on
Monograph No. 11 USBR 1952 Performance-Based Design in Earthquake
Geotechnical Engineering. Taormina, Italy.


ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017
Appendix A

Intensity Effects
I Not felt. Marginal and long period effects of large earthquakes.
II Felt by persons at rest, on upper floors, or favourably placed.
III Felt indoors. Hanging objects swing. Vibration like passing of light trucks. Duration
estimated. May not be recognised as an earthquake.
IV Hanging objects swing. Vibration like passing of heavy trucks or sensation of a jolt
like a heavy ball striking the walls. Standing motor cars rock. Windows, dishes,
doors rattle. Glasses clink. Crockery clashes. In the upper range of IV wood walls
and frames creak.
V Felt outdoors, duration estimated. Sleepers wakened. Liquids disturbed, some spilled.
Small unstable objects displaced or upset. Doors swings, close, open. Shutters,
pictures move. Pendulum clocks stop, start, change rate.
VI Felt by all. Many frightened and run outdoors. Persons walk unsteadily. Windows,
dishes, glassware broken. Knickknacks, books, etc. off shelves. Pictures off walls.
Furniture moved or overturned. Weak plaster and masonry D cracked. Small bells
ring (church, school). Trees, bushes shaken (visibly, or heard to rustle – CFR).

VII Difficult to stand. Noticed by drivers of motor cars. Hanging objects quiver.
Furniture broken. Damage to masonry D, including cracks. Weak chimneys broken at
roof line. Fall of plaster, loose bricks, stones, tiles, cornices (also unbraced parapets
and architectural ornaments – CFR). Some cracks in masonry C. Waves on ponds;
water turbid with mud. Small slides and caving in along sand or gravel banks. Large
bells ring. Concrete irrigation ditches damaged.
VIII Steering of motor cars affected. Damage to masonry C, partial collapse. Some
damage to masonry B, none to masonry A. Fall of stucco and some masonry walls.
Twisting, fall of chimneys, factory stacks, monuments, towers, elevated tanks. Frame
houses moved on foundations if not bolted down; panel walls thrown out. Decayed
piling broken off. Branches broken from trees. Changes in flow or temperature of

springs and wells. Cracks in wet ground and on steep slopes.

IX General panic. Masonry D destroyed; masonry C heavily damaged, sometimes with
complete collapse; masonry B seriously damaged. (General damage to foundations –
CFR). Frame structures, if not bolted, shifted off foundations. Frames cracked.

Serious damage to reservoirs. Underground pipes broken. Conspicuous cracks in

ground. In alluviated areas sand and mud ejected, earthquake fountains, sand craters.
X Most masonry and frame structures destroyed with their foundations. Some well-built
wooden structures and bridges destroyed. Serious damage to dams, dikes,
embankments. Large landslides. Water thrown on banks of canals, rivers, lakes, etc.
Sand and mud shifted horizontally on beachheads and flat land. Rails bent slightly.
XI Rails bent greatly. Underground pipelines completely out of service.
XII Damage nearly total. Large rock masses displaced. Lines of sight and level distorted.
Objects thrown into the air.
Note: Masonry A, B, C, D. To avoid ambiguity of language, the quality of masonry, brick or otherwise, is specified
by the following lettering (which has no connection with the conventional Class A, B, C construction).
– Masonry A: Good workmanship, mortar, and design; reinforced, especially laterally, and bound together by using
steel, concrete, etc.; designed to resist lateral forces;
– Masonry B: Good workmanship and mortar; reinforced, but not designed to resist lateral forces;
– Masonry C: Ordinary workmanship and mortar; no extreme weaknesses such as non-tied-in corners, but masonry
is neither reinforced nor designed against horizontal forces;
– Masonry D: Weak materials, such as adobe; poor mortar; low standards of workmanship; weak horizontally;
– CFR indicates additions to classification system by Richter (1958).

ANCOLD Guidelines for Design of Dams and Appurtenant Structures for Earthquake – Draft March 2017