You are on page 1of 61

0068_C07_fm Page 1 Wednesday, July 31, 2002 8:20 AM

7
Geotechnical and
Foundation Aspects

7.1 Introduction
7.2 Seismic Hazards
7.3 Strong Ground Motion
Characterization of Strong Ground Motion · Site-Specific Soil
and Topographic Effects · Design Ground Motions
7.4 Dynamic Soil Behavior
7.5 Liquefaction
General Concepts · Magnitude Scaling Factors and Other
Corrections · Liquefaction Settlement · Shear Strength of
Liquefied Soil
7.6 Seismic Analysis of Slopes and Dams
7.7 Earthquake-Resistant Design of Retaining Walls
7.8 Soil Remediation Techniques for Mitigation
of Seismic Hazards
Defining Terms
Horst G. Brandes References
University of Hawaii, Honolulu Further Reading

7.1 Introduction
Most earthquakes involve a sudden release of energy as rock masses in the Earth’s crust fracture due to
tectonic faulting. Some are also associated with volcanic processes, landslides, large explosions, or even
ground loading from the filling of large reservoirs. Following an earthquake, energy propagates outward
from the source in the form of elastic seismic waves . The duration, amplitude, and frequency charac-
teristics of these waves are a function of the type and magnitude of the earthquake, the distance from
the epicenter , and the type and distribution of geologic materials through which the waves travel.
Structures located along the way are subject to shaking and may experience damage or even fail cata-
strophically. Whereas the broad field of earthquake engineering deals with all of these aspects, the
perspective of geotechnical engineers is often more site-specific, dealing with issues relating to the
propagation of seismic waves through soil and rock layers close to the ground surface, the effects that
shaking has on ground displacements and stresses, as well as on civil engineering structures such as dams,
slopes, retaining walls, foundations, and lifelines. A proper understanding of geotechnical earthquake
engineering principles is necessary for rational and sound earthquake-resistant design. The focus of this
chapter is on geotechnical aspects. The reader is directed to the rest of this volume for additional topics
on earthquake engineering.
Geologists, seismologists, and geophysical scientists have addressed earthquakes in earnest for at least
100 years, although they have been concerned primarily with low-intensity seismic effects. Geotechnical
earthquake engineering is a comparatively younger discipline. Significant progress has only been made

© 2003 by CRC Press LLC


0068_C07_fm Page 2 Wednesday, July 31, 2002 8:20 AM

7-2 Earthquake Engineering Handbook

60°N

40°N

20°N

150°W 120°W 90°W


FIGURE 7.1 National Strong Motion Program recording stations as of April 10, 2002. (Courtesy National Strong
Motion Program, U.S. Geological Survey)

in the past 30 to 40 years, especially after the 1971 San Fernando earthquake, when large numbers of
strong motion sensors were installed to record high-magnitude earthquakes of the type that can cause
damage to civil engineering structures. Understanding of the effects of strong ground shaking is now
advancing rapidly because of the placement of large networks of digital sensors in seismically active
regions of the world, including the United States, Japan, Taiwan, Europe, and elsewhere. Data sets from
earthquakes outside the United States are becoming increasingly more accessible.
The U.S. Geological Survey (USGS) operates the National Strong Motion Program (http://
nsmp.wr.usgs.gov), which oversees a network of more than 900 instruments located at approximately
628 permanent stations throughout North America, Hawaii, and the Caribbean (Figure 7.1).
Records of strong motion can be obtained from the USGS for selected earthquakes dating back as far
as 1933. Strong motion records are also available from other government agencies, such as the California
Geological Survey (http://docinet3.consrv.ca.gov/csmip/) and the National Oceanic and Atmospheric
Administration (http://www.ngdc.noaa.gov/seg/hazard/strong.html). Various universities and research
institutions also maintain strong motion databases, including the University of California at Berkeley
(http://peer.berkeley.edu/smcat/), the University of Washington (http://www.geophys.washington.edu/SEIS/
PNSN/SMO/), the Lamont-Doherty Observatory (http://www.Ideo.columbia.edu/nceer/nceer.html),
and the University of California at Santa Barbara (http://db.cosmos-eq.org/), among others.
A number of recent earthquakes have provided an unprecedented wealth of new data that are currently
the focus of intense study. This has already resulted in a better understanding of how local soil conditions
affect ground motions, liquefaction, and the dynamic behavior of structures. Additional developments
are certain to follow in the years to come [Finn, 2000].

7.2 Seismic Hazards


Earthquakes can be tremendously deadly and expensive, with large economic and social effects. Earth-
quake engineering is concerned with correctly identifying and minimizing seismic hazards and risk

© 2003 by CRC Press LLC


0068_C07_fm Page 3 Wednesday, July 31, 2002 8:20 AM

Geotechnical and Foundation Aspects 7-3

FIGURE 7.2 Building that collapsed during the 1999 Turkey (Kocaeli) earthquake due to liquefaction of the sub-
surface with little or no damage to the structure itself. (Courtesy J. P. Bardet)

FIGURE 7.3 Building heavily damaged during the 1999 Taiwan (Chi-Chi) earthquake. Intense shaking caused the
failure of load-bearing columns in the lower floors. (Courtesy J. P. Bardet)

through increased public awareness, sensitive land development policies, and implementation of techni-
cally sound seismic building codes. Past earthquakes have shown that the effects of strong ground shaking
are diverse but can broadly be divided into two classes depending on whether they lead to permanent
ground displacements or not. Damage to constructed structures can occur in either case. The distinction
between the two types of distress is illustrated in Figures 7.2 and 7.3. The first image shows a building
that collapsed during the 1999 Turkey (Kocaeli) earthquake due to liquefaction of the subsurface, with

© 2003 by CRC Press LLC


0068_C07_fm Page 4 Wednesday, July 31, 2002 8:20 AM

7-4 Earthquake Engineering Handbook

FIGURE 7.4 Ground cracks that developed during the July 1990 Philippines earthquake. (Courtesy EQE International)

little or no damage to the structure itself. In contrast, Figure 7.3 shows a building that was heavily
damaged due to the collapse of a portion of the two lower floors during the 1999 Taiwan (Chi-Chi)
earthquake. In this case, there was no evidence of large residual ground displacements. Intense shaking
caused the failure of load-bearing columns in the lower floors.
Among the hazards associated with permanent ground displacements are the development of ground
cracks, surface faults, liquefaction-induced settlements and lateral spreads, ground heave, slope and rock
failures, and settling and failure of earth structures. Ground cracks can range from small fractures to
long and deep depressions that in some cases can extend over long distances (Figure 7.4).
Where these cracks transect roads, railroads, dams, or populated centers, they can cause widespread
damage (Figures 7.5 and 7.6). If water fills these surface fractures, the resulting lateral pressures can lead
to failures of slopes, embankments, and retaining structures. Faulting involves the vertical offset of the
ground surface, which in some cases can be on the order of several meters (Figures 7.7 to 7.9). The
formation of ground cracks and faults as a result of earthquakes is difficult to predict because their
distribution and extent is highly dependent on geologic and soil nonconformities, which are inherently
random. No reliable design procedures exist that account for these types of failures.
Liquefaction, i.e., the temporary loss of soil strength and fluidization that occurs in certain saturated
granular soils due to seismic shaking, represents a significant hazard in coastal areas and other locations
with a high water table. The reduction in bearing capacity that accompanies the process of liquefaction
can lead to the sinking of buildings, bridges, and other heavy structures, often with little or no damage
to the structure itself (Figure 7.2). Liquefied beds can sometimes be detected by the presence of surface
sand boils, formed as mixtures of soil and water squeeze out of the ground, with the soil residue deposited
nearby (Figure 7.10).
Significantly more hazardous are flow failures that involve the lateral displacement of fluidized, or
nearly fluidized, soil under the effect of gravity. Such lateral spreading can take place at ground

© 2003 by CRC Press LLC


0068_C07_fm Page 5 Wednesday, July 31, 2002 8:20 AM

Geotechnical and Foundation Aspects 7-5

FIGURE 7.5 Ground cracking transecting roadway, 1991 Costa Rica earthquake. (Courtesy EQE International)

FIGURE 7.6 Cracks up to 6 m deep devel-


oped on the upstream face of Fatehgadh Dam
during the 2001 India (Bhuj) earthquake.
(Courtesy J. P. Bardet)

inclinations of as little as 0.1% and can result in deformations ranging from a few millimeters to tens of
meters [Bardet et al., 1999]. They can cause damage to dams (Figure 7.11), harbor structures
(Figure 7.12), buildings (Figure 7.13), estuarine and river embankments (Figure 7.14), bridges, utilities,
roadways, and airport runways. Liquefied deposits that do not undergo significant lateral displacements
can still settle a considerable amount due to subsequent consolidation, and therefore can cause damage
as well. The engineering treatment of liquefaction has progressed significantly over the past 10 years and

© 2003 by CRC Press LLC


0068_C07_fm Page 6 Wednesday, July 31, 2002 8:20 AM

7-6 Earthquake Engineering Handbook

FIGURE 7.7 Failure of the Shihkang Dam during the 1999 Taiwan (Chi-Chi) earthquake due to tectonic uplifting
and compression associated with a large fault transecting the structure. (Courtesy J. P. Bardet)

FIGURE 7.8 Fault intersecting running track at local school in Wu Feng, with an offset of approximately 1.7 m,
due to the 1999 Taiwan (Chi-Chi) earthquake. (Courtesy J. P. Bardet)

can now be considered a subdiscipline in its own right [Seed et al., 2001]. It is addressed in building
codes and is currently the subject of extensive research.
Earthquakes can cause the failure of soil and rock slopes, particularly those that are marginally stable
to begin with. The most common type of landslides triggered by seismic events include rock falls, soil
slides, and rock slides on relatively steep slopes that are covered with disaggregated soil and rock
[Wieczorek, 1996]. The debris from such failures can cut off roads and streams, and it can damage
buildings, bridges, and other structures (Figures 7.15 and 7.16).
Water that is impounded behind streams that are blocked by debris can break through suddenly and
cause additional destruction. Loss of life due to landslides is a common occurrence. Of special concern

© 2003 by CRC Press LLC


0068_C07_fm Page 7 Wednesday, July 31, 2002 8:20 AM

Geotechnical and Foundation Aspects 7-7

FIGURE 7.9 Fault rupture across road led to a lateral offset of 4 m during the 1999 Turkey (Duzce) earthquake.
(Courtesy J. P. Bardet)

FIGURE 7.10 Sand boils showing vents and sand and silt deposits from spouting associated with liquefied beds,
1976 Guatemala earthquake. (U.S. Geological Survey photo, courtesy National Oceanic and Atmospheric Adminis-
tration-National Geophysical Data Center)

© 2003 by CRC Press LLC


0068_C07_fm Page 8 Wednesday, July 31, 2002 8:20 AM

7-8 Earthquake Engineering Handbook

FIGURE 7.11 Liquefaction failure of Lower San Fernando Dam during the 1971 San Fernando earthquake. Analysis
of this particular failure led to significant advances in the understanding of liquefaction phenomena. (K. Steinbrugge
Collection, courtesy Earthquake Engineering Research Center, University of California, Berkeley)

FIGURE 7.12 Lateral spreading and sand boils in fill material retained by pier structure, 1995 Japan (Kobe) earth-
quake. (Kobe Geotechnical Collection, courtesy Earthquake Engineering Research Center, University of California,
Berkeley)

© 2003 by CRC Press LLC


0068_C07_fm Page 9 Wednesday, July 31, 2002 8:20 AM

Geotechnical and Foundation Aspects 7-9

FIGURE 7.13 Subsidence and lateral spreading due to liquefaction lead to the collapse of buildings along the
shoreline in the fishing village of Guzelyali during the 1999 Turkey (Kocaeli) earthquake. (J.C. Borrero photo, courtesy
National Oceanic and Atmospheric Administration-National Geophysical Data Center)

FIGURE 7.14 This sidewalk slipped into the water due to the combined effects of lateral spreading, subsidence, and
a tsunami that resulted from the 1999 Turkey (Kocaeli) earthquake. (J.C. Borrero photo, courtesy National Oceanic
and Atmospheric Administration-National Geophysical Data Center)

are slope failures that progress quickly, where the displaced material moves at speeds in excess of 5 m/sec.
This is approximately the velocity at which a person can run [Cruden and Varnes, 1996]. A tragic example
is the Huascaran Mountain landslide that occurred as a result of the 1970 (M 7.7) Peru earthquake. It
buried the town of Yungay and part of the town of Ranrahirca, with a loss of life exceeding 18,000. In
many earthquakes, landslides are the principal cause of death and destruction, and therefore constitute
the greatest hazard. It is not uncommon for a single earthquake to trigger thousands of landslides,
although most are usually small and take place in remote areas. The stability of slopes under seismic
conditions can be analyzed by a number of techniques, ranging from pseudostatic to fully dynamic

© 2003 by CRC Press LLC


0068_C07_fm Page 10 Wednesday, July 31, 2002 8:20 AM

7-10 Earthquake Engineering Handbook

FIGURE 7.15 A large landslide triggered during the 1999 Taiwan (Chi-Chi) earthquake altered the path of the
Taichai River, forcing it onto adjacent agricultural land. (Courtesy J. P. Bardet)

FIGURE 7.16 Numerous rock slides occurred along roads cut into mountain sides, involving both colluvium and
bedrock, during the Southern Peru earthquake of 2001. (Courtesy Joseph Wartman, Drexel University)

numerical methods. Somewhat more problematic but equally important is estimating the runoff distance
of the waste debris.
Retaining walls and earth structures are susceptible to failure either due to liquefaction of fill material
or due to severe shaking leading to dynamic lateral pressures that exceed the capacity of the soil-retaining
structure system (Figures 7.12, 7.17, and 7.18). If located in seismically active areas, harbor structures
such as gravity quay walls, sheet pile bulkheads, and pile and caisson piers must be designed for seismic
loading, especially if their failure carries unacceptable risk of loss of life or dire economic consequences.
Design principles for retaining walls and earth fill dams are described in this chapter.
Lifelines consist of infrastructure systems whose purpose is to deliver electric power, water, sewage,
gas, oil, telephone, and digital data. Many of these utilities are installed below ground and can be damaged

© 2003 by CRC Press LLC


0068_C07_fm Page 11 Wednesday, July 31, 2002 8:20 AM

Geotechnical and Foundation Aspects 7-11

FIGURE 7.17 Extensional cracks up to 12 cm wide developed in this retaining spillway wall of the Austrian Dam
in the Santa Cruz Mountains of California as a result of the 1989 Loma Prieta earthquake. (G.R. Fisher photo,
courtesy U.S. Geological Survey)

FIGURE 7.18 Extensive damage occurred to the Kobe waterfront due to widespread ground failure associated with
liquefaction-related phenomena during the 1995 Japan (Kobe) earthquake. (R. Hutchinson photo, courtesy National
Oceanic and Atmospheric Administration-National Geophysical Data Center)

by ground movement. In addition to utilities, lifelines also include transportation systems such as roads,
railroads, airports, and harbors. In a modern city, lifelines constitute complex systems that are crucial
for safety and economic reasons. Their interruption due to earthquakes can be devastating. Much recent
research is focusing on ways to minimize hazards to lifelines from earthquakes.
Another significant hazard associated with certain earthquakes is the generation of tsunamis, which
are long water waves that form when the seafloor displaces suddenly. Tsunamis are a concern primarily
in the Pacific Ocean due to the large number of earthquakes that occur along its rim. However, not all
earthquakes lead to tsunamis. It appears that only those that involve dip-slip faulting or seabed mass
wasting lead to tsunamis. History includes numerous examples of destructive events. For example, in

© 2003 by CRC Press LLC


0068_C07_fm Page 12 Wednesday, July 31, 2002 8:20 AM

7-12 Earthquake Engineering Handbook

1960 an M 9.5 earthquake in Chile generated a large tsunami that resulted in the deaths of 231 people
in Hawaii, Japan, the Phillipines, and the west coast of the United States. More recently, the 1998 Papua
New Guinea M 7.0 earthquake caused a massive underwater slump that led to a large tsunami, which
destroyed several villages and killed more than 2000 people in the Sissano area [Tappin et al., 1999; Geist,
2000]. Maximum wave heights were estimated at 15 m [Geist, 2000].

7.3 Strong Ground Motion

7.3.1 Characterization of Strong Ground Motion


Seismic waves result in complex ground displacements that vary from one location to the next. Seis-
mometers are instruments that allow recording of the time-history of this motion. Of interest to engineers
are devices that are designed to record the type of strong ground motions that can cause damage in an
earthquake. Modern instruments use sets of accelerometers to measure components of acceleration in
three orthogonal directions, two in the horizontal direction and one in the vertical (Figure 7.19). Although
analog systems that record data on photographic film are still in widespread use, they are rapidly being
replaced and supplemented with digital systems that collect data electronically at rates of tens or hundreds
of measurements per second. These devices convert the analog transducer data into digital data and store
the information on solid-state memory. Some instruments can be accessed remotely and the data down-
loaded through telemetry. Seismometers remain in standby mode until they are triggered by motion
above a certain threshold.
The raw digital data obtained by strong motion recorders must be corrected for various type of error,
including the instrument’s intrinsic dynamic response, background noise, and baseline errors associated
with the level of motion necessary for triggering the instrument. Software is available from the USGS to
correct digitized records of strong motion earthquakes. The computer program BAP (Basic Strong
Motion Accelerogram Processing Software) can be used to apply linear baseline corrections, instrument
corrections, and to filter high and low frequency components from the record. It is described in more
detail and is available on the Internet for download at http://nsmp.wr.usgs.gov/processing.html.
An acceleration time series can be integrated numerically to obtain the corresponding velocity and
displacement time series (Figure 7.20). Time histories, such as those in Figures 7.19 and 7.20, represent
the complete record of motion at the instrument site. From a geotechnical point of view, the duration,
frequency content, and amplitude of the motions are important. A number of parameters have been
proposed to express these characteristics; the most commonly used is peak ground acceleration (PGA),
which represents the largest recorded acceleration. It can be determined directly from the accelerogram
and is typically expressed in fractions of gravity g, where 1 g is equal to 9.81 m/sec2. The focus is usually
on peak horizontal accelerations (PHA) (sometimes the average of the two orthogonal PHAs), due to
their role in determining lateral inertial forces in structures. However, peak vertical horizontal acceler-
ations (PVA) may be of importance when considering the effects of earthquakes on light structures,
whose weight may be insufficient to resist vertical dynamic loads from strong shaking. For practical
purposes, the PVA has often been taken to be two thirds of the PHA [Newark and Hall, 1982]. For
example, the records in Figure 7.19 indicate a PHA of 0.358 g and a PVA of 0.229 g, for a ratio of 0.64.
However, recent research has shown that this ratio approaches unity close to the earthquake fault (see
Chapter 5, this volume). As a general rule of thumb, acceleration levels must approach 0.1 g to produce
damage in the weakest of constructions, whereas accelerations in excess of 0.5 g are deemed to be very
dangerous to many structures [Arnold and Reitherman, 1982]. Of course, duration and frequency are
of importance as well. Peak horizontal accelerations have been correlated to earthquake intensity through
the Modified Mercalli Intensity Scale, which rates the effects of an earthquake on humans, structures,
and the ground at specific locations (Figure 7.21; see also Chapter 4 of this volume).
Peak horizontal velocity (PHV) and peak horizontal displacement (PHD) are sometimes reported.
They each reflect different sensitivities to the frequency components of the motion. In some cases,
accelerations and velocities refer not to peak values but to some other level of motion. For example,

© 2003 by CRC Press LLC


0068_C07_fm Page 13 Wednesday, July 31, 2002 8:20 AM

Geotechnical and Foundation Aspects 7-13

0.4
180 Degrees

Acceleration (g)

0.0

-0.4
0 5 10 15 20 25 30
Time (s)

0.4
270 Degrees
Acceleration (g)

0.0

-0.4
0 5 10 15 20 25 30
Time (s)

0.4
Vertical
Acceleration (g)

0.0

-0.4
0 5 10 15 20 25 30
Time (s)

FIGURE 7.19 Accelerogram of recorded motions along three mutually perpendicular directions, obtained during
the 1999 Turkey (Kocaeli) earthquake (M 7.4).

terms such as effective, sustained maximum, and effective design accelerations or velocities are occasion-
ally used [Kramer, 1996].
Earthquake-induced ground motions contain a broad range of frequencies. The dynamic response of
compliant structures is highly dependent on the frequency characteristics of these motions, which can
be expressed conveniently with reference to a linear single-degree-of-freedom (SDOF) system, as shown
in Figure 7.22. Displacement of the ground is noted by ug(t), and the relative displacement between the

© 2003 by CRC Press LLC


0068_C07_fm Page 14 Wednesday, July 31, 2002 8:20 AM

7-14 Earthquake Engineering Handbook

80.0
180 Degrees

Velocity (cm/s)

0.0

-80.0
0 5 10 15 20 25 30
Time (s)

80.0
180 Degrees
Displacement (cm)

0.0

-80.0
05 10 15 20 25 30
Time (s)

FIGURE 7.20 Velocity and displacement time histories corresponding to the 180˚ accelerogram in Figure 7.19.

FIGURE 7.21 MMI scale. (From Trifunac, M.D. and A.G. Brady. 1975. “On the Correlation of Seismic Intensity
with Peaks of Recorded Strong Ground Motion,” Bull. Seismol. Soc. Am., 65, 139–162. With permission.)

© 2003 by CRC Press LLC


0068_C07_fm Page 15 Wednesday, July 31, 2002 8:20 AM

Geotechnical and Foundation Aspects 7-15

ug(t)

c u(t)
m
k

FIGURE 7.22 Single-degree-of-freedom system.

ground and the mass m by u(t). Hence the total, i.e., absolute, displacement of the mass becomes
uT(t) = u(t) + ug(t). The response of the SDOF system is given by the following equation of motion:

( )
m u˙˙ (t ) + u˙˙g (t ) + cu˙ (t ) + ku (t ) = 0 (7.1)

where c is the viscous damping coefficient of the dashpot and k is the spring stiffness. The first term on
the left-hand side represents the inertial force, the second term the damping force, and the third term
the elastic force. Equation 7.1 can be rewritten as:

( )
2 ηωu˙ (t ) + ω 2u (t ) = − u˙˙ (t ) + u˙˙g (t ) (7.2)

where ω is the resonant natural frequency of the system, ω = k / m = 2π / T , T is the resonant period,
and η is the damping ratio expressed as a fraction of the critical damping, η = c/ccr = c/(2m ω). The
ground acceleration u g(t ) would be given by site-specific acceleration time histories, such as the ones in
Figure 7.19. Solution of Equation 7.2, which for the case of complex ground motions ug(t ) must be
accomplished numerically, yields the relative displacement history u (t ). Notice that the displacement is
a function of time t, resonant frequency ω, and damping ratio η.
The response spectrum is a plot of the maximum computed displacement as a function of natural
frequency of the SDOF system:

Sd (ω, η) = max u (t , ω, η) (7.3)

Spectra are determined for a given level of damping, typically 5%. In a similar fashion, velocity u̇(t ) and
acceleration u˙˙ (t) spectra can be determined by numerical differentiation of the displacement u(t).
However, in practice it is more common to make use of pseudovelocity (Sv or PSV) and pseudoacceleration
(Sa or PSA) spectra, which are computed directly from Sd by noting that the three quantities are related:

Sv = ω Sd (7.4)

Sa = ω 2Sd (7.5)

Sv and Sd , which are different in magnitude from the spectra based on u̇(t) and u˙˙ (t), are more meaningful
quantities. Sv can be related to the peak value of strain energy stored in the system during a given
earthquake, and Sd can be related to the base shear coefficient used in building codes to estimate base
shear forces in structures.
Sd , Sv , and Sa spectra are shown in Figure 7.23 for the 180° accelerograph in Figure 7.19. The three
spectra illustrate how displacement, velocity, and acceleration are each associated with different frequency
ranges. Peak displacements occur at comparatively high periods (or low frequencies), whereas peak
accelerations are found at low periods (or high frequencies). Stiff structures with a low resonant period
are most sensitive to the low period components of the ground motion, therefore consideration of the
acceleration spectrum is of critical importance. On the other hand, flexible structures with a high resonant

© 2003 by CRC Press LLC


0068_C07_fm Page 16 Wednesday, July 31, 2002 8:20 AM

7-16 Earthquake Engineering Handbook

150
5% Damping

100

Sd (cm) 50

0
0 5 10 15 20
Period (s)

160
5% Damping
Sv or PSV (cm/s)

120

80

40

0
0 5 10 15 20
Period (s)

1.2
5% Damping
Sa or PSA (g)

0.8

0.4

0.0
0 5 10 15 20
Period (s)

FIGURE 7.23 Response spectra for 180° accelerograph in Figure 7.19.

period are more sensitive to the high-period components of the motion, which are better reflected in
the displacement spectrum. Because Sd , Sv , and Sa are related, they are sometimes combined in a tripartite
plot that uses four logarithmic scales to display displacement, velocity, acceleration, and either period
or frequency (Figure 7.24).
Linear response spectra are widely used in earthquake engineering as a practical means of character-
izing strong ground motions. However, it should be remembered that most structures behave inelastically
when subjected to strong ground motions. Inelastic response spectra can be determined to account for
nonlinearities in the force-displacement relationship [Newark and Hall, 1982]. The extent to which a
particular structure behaves inelastically can be defined by a ductility factor µ:

up
µ= (7.6)
uy

© 2003 by CRC Press LLC


0068_C07_fm Page 17 Wednesday, July 31, 2002 8:20 AM

Geotechnical and Foundation Aspects 7-17

10
0,
00
0
10
00

,0
,0

00
10
2
10

)
2
/s
m
10

(c
00

00

A
10

PS
S
d
(c
m
)

10
0
10

0
1
10
PSV (sec)

10
10

1
0
10

1
0.

0.
1
0.

01
01

0.
10
0.

1
00

00
1

0.

01
00
0.
1 0 1 2
10 10 10 10

ω (rad/sec)

FIGURE 7.24 Tripartite plot of response spectral displacement, velocity, and acceleration for the record in
Figure 7.23.

where u p is the peak displacement of the mass for a given ground excitation and uy is the displacement
of the mass necessary to yield inelastic response. Inelastic response spectra can be calculated in the time
domain by assuming that the stiffness term in Equation 7.1 is a function of displacement. When compared
to the elastic case, inelastic spectral accelerations decrease with ductility. This topic is covered in Chapter 4
of this volume, as well as in Chopra [2001].
The duration of strong ground motion, and in particular the number of load reversals, has a strong
influence on the damage that a particular earthquake can cause. A moderate earthquake of long duration
can result in equal or larger destruction than a short, intense one. Structures that are loaded beyond a
cyclically stable range may fail after a given number of load reversals. The same is true for foundation
elements that may lose their support upon repeated load reversals due to reduction in soil stiffness and
strength. Also, pore pressures in saturated, loose granular soils can accumulate with load cycles and lead
to liquefaction after a given number of load reversals. Duration of the strong motion component of a
seismic event can be determined from an accelerogram using one of a number of criteria [Kramer, 1996].
The most straightforward approach is to measure the time between the first and last exceedance of some
threshold acceleration, typically 0.05 g. This is referred to as the bracketed duration [Bolt, 1969].

7.3.2 Site-Specific Soil and Topographic Effects


In general, ground motions decrease with distance from the earthquake’s epicenter. Ground motion
attenuation is a function of the type and magnitude of the earthquake, distance from the source, and
the geology through which seismic waves travel. Attenuation relationships are crucial for determining
earthquake hazard assessments and for establishing seismic design criteria for engineered structures.

© 2003 by CRC Press LLC


0068_C07_fm Page 18 Wednesday, July 31, 2002 8:20 AM

7-18 Earthquake Engineering Handbook

Attenuation is usually expressed in terms of peak ground acceleration (PGA), peak ground velocity
(PGV), or spectral ground acceleration (PSA), at 5% damping. For example, the 1996 U.S. National
Seismic Hazard maps prepared by the USGS display PGA and PSA at periods of 0.2, 0.3, and 1.0 sec,
with a 10%, 5%, and 2% probability of exceedance in 50 years (http://geohazards.cr.usgs.gov/eq/).
A number of attenuation relationships have been developed over the years and new data from recent
earthquakes will undoubtedly lead to refinements in the future. These expressions require input param-
eters that describe earthquake magnitude, type of faulting, distance, and local soil or rock characteristics.
Specific ground motion attenuation relationships are associated with particular tectonic environments
and are derived empirically by statistical means from large numbers of earthquakes or synthetically by
assuming a particular seismological model. Relationships have been developed for shallow crustal earth-
quakes in active tectonic regions such as western North America, for shallow earthquakes in stable
continental regions such as central and eastern North America, and for earthquakes in subduction zones
such as Alaska [Abrahamson and Shedlock, 1997]. A detailed description of current ground motion
attenuation relationships in use in North America can be found in a 1997 special issue of Seismological
Research Letters (Vol. 68, No. 1) and in Chapter 5. Site-specific conditions are usually included only in
crude terms through broad classifications such as rock/stiff soil, deep soil, and soft soil. A notable
exception is the model of Boore et al. [1997], which takes into account the average shear wave velocity
to a depth of 30 m. Attenuation relationships have also been developed for use in Japan and elsewhere
[Molas and Yamazaki, 1995, 1996; Shabestari and Yamazaki, 1998].
It has long been recognized that soils whose stiffness differs from that of the underlying bedrock can
affect the amplitude, frequency, and duration content of surface strong motions. In particular, thick soft
soil deposits have been observed to cause amplification of ground motions. Important evidence was
collected during the 1985 Michoacan (Mexico) earthquake (M8.1) and during the 1989 Loma Prieta
earthquake (MW 7.1). The center of Mexico City is partially located on soft lake clays that extend down
50 m or more. Further to the west, the clay deposits thin out and are gradually replaced by granular and
rock materials of volcanic origin. Figure 7.25 illustrates how spectral accelerations at the SCT site,
overlying approximately 38 m of clay, are significantly larger than at the UNAM site, which was located
on basaltic rock and hard soil. The frequency content is also distinctly different. Whereas spectral energy
is concentrated around a period of 2 sec at the SCT site, the UNAM spectrum is much broader. Severe
damage to buildings occurred where the clay thickness exceeded 40 m and where the natural period of
the soil deposit matched that of the structure [Stone et al., 1987]. Amplification of ground motions at

5% Damping

SCT
Sa (g)

UNAM

T (s)

FIGURE 7.25 Spectral accelerations at SCT site (soft soil) vs. UNAM site (rock), 1985 Mexico earthquake. (From
Romo, M.P. and H.B. Seed. 1986. “Analytical Modeling of Dynamic Soil Response in the Mexico Earthquake of
September 19, 1985,” in Proc. ASCE International Conference on the Mexico Earthquakes 1985, Mexico City, American
Society of Civil Engineers, New York. With permission.)

© 2003 by CRC Press LLC


0068_C07_fm Page 19 Tuesday, August 13, 2002 1:55 PM

Geotechnical and Foundation Aspects 7-19

sites with thick soil deposits was also observed during the 1989 Loma Prieta (M7.1) earthquake. In
general, seismic shaking decreased with distance from the epicenter, located in the Santa Cruz Mountains,
but areas underlain by substantial deposits of alluvium and bay mud clearly suffered the worst damage.
For example, local site conditions at the northern section of the Cypress viaduct, consisting of engineered
fill and soft to medium stiff bay mud, strongly amplified peak ground surface accelerations, and especially
the long-period components, causing the collapse of the bridge in that area. On the other hand, the
southern section was not founded on bay mud and did not collapse [Seed et al., 1990].
Based in large part on extensive measurements taken during the 1985 Michoacan and the 1989 Loma
Prieta earthquakes, Idriss [1990] has proposed an approximate relationship between peak accelerations
on rock sites vs. peak accelerations on soil sites (Figure 7.26). Seed et al. [1976] considered the effects of
different soil types on average spectral acceleration (Figure 7.27). It is clear from Figure 7.27 that the
presence of soil also affects the frequency content of ground motions. In particular, soft and deep deposits
result in a shift towards higher periods with significant energy at values of 1 sec and higher. Such periods
may be similar to those of large structures and therefore they can induce dangerous inertial forces.
A simple analytical model can be used to illustrate some of the important aspects of wave propagation
through surface soil layers. Assume a uniform deposit, consisting of a viscoelastic soil of thickness H,
located above a rigid bedrock (Figure 7.28). The equation of motion describing harmonic shear waves
traveling upward from the bedrock can be solved to obtain the horizontal soil displacement through a
soil layer having a shear velocity Vs [Kramer, 1996]:

ω i ωt
u ( z , t ) = A cos ze (7.7)
Vs

This equation represents a standing wave with amplitude A cos(ω/Vs ). The ratio of maximum displace-
ment at the ground surface to maximum displacement at the soil–bedrock interface is given, for small
damping ratios η, by:

1
A= (7.8)
cos (ωH Vs ) + ( ηωH Vs )
2 2

0.6

0.5
Acceleration - soil sites (g)

0.4

0.3

0.2

0.1

0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6
Acceleration - rock sites (g)

FIGURE 7.26 Peak accelerations on rock sites vs. peak accelerations on soil sites. (From Idriss, I.M. 1990. H. Bolton
Seed Memorial Symposium, BiTech Publishers. With permission.)

© 2003 by CRC Press LLC


0068_C07_fm Page 20 Tuesday, August 13, 2002 1:55 PM

7-20 Earthquake Engineering Handbook

Soft to medium clay and sand (15 records)

Deep cohesionless soils, >250ft (30 records)

Stiff soils, <200ft (31 records)

Rock (28 records)

FIGURE 7.27 Effect of soil type on average spectral acceleration at 5% damping. (From Seed, H.B., C. Ugas et al.
1976. “Site-Dependent Spectra for Earthquake-Resistant Design,” Bull. Seismol. Soc. Am., 66, 221–243. With permission.)

Ground surface

Soil Deposit
with shear z
H
wave velocity
Vs

Rigid Bedrock

FIGURE 7.28 Simple model for vertical wave propagation.

Note that amplification is a function of the frequency of the motion and the damping of the soil.
Amplification maxima occur at the following frequency values:

Vs  π 
ωn =  + n π (7.9)
H 2 

where n = 0, 1, 2,… These are the natural or resonant frequencies of the soil deposit. For an undamped
soil, they result in infinite amplification.
At finite damping values, amplification decreases with increasing frequency and the maximum occurs
at the lowest natural frequency, i.e., at n = 0. This is known as the fundamental frequency, ω0 = πVs /(2H).
The characteristic site period corresponding to this frequency is T0 = 4H/Vs. For example, assuming that
Vs = 450 m/sec, H = 150 m, and η = 5%, the characteristic site period becomes T0 = 1.34 sec, and the

amplification factor A = 12.7 (Figure 7.29). This means that ground motion with a concentration of
energy near this period will result in resonance and more than a tenfold amplification of horizontal ground
displacement. If we use the rule of thumb that the fundamental period of a building with N floors is
approximately N/10, then buildings with 13 or 14 floors would be in resonance and would face the potential
of severe damage if not properly designed and built. For level stratified deposits with approximately
horizontal soil layers, shear velocities can be averaged as a first approximation. Similar methods of analysis

© 2003 by CRC Press LLC


0068_C07_fm Page 21 Tuesday, August 13, 2002 1:55 PM

Geotechnical and Foundation Aspects 7-21

15
η=0

Amplification factor 10
η = 5%
5

0
0 5 10 15 20 25 30
Natural frequency, ω

FIGURE 7.29 Site amplification as a function of frequency and soil damping for VS = 450 m/sec and H = 150 m.

for one-dimensional wave propagation are presented by Kramer [1996] for uniform and for layered soil
on elastic rock.
The assumptions made in these types of analysis are quite simplistic. Most soils do not behave elastically
and their properties can vary significantly over short distances. Propagation of seismic waves through
soil layers invariably involves some level of material nonlinearity. At all but the smallest levels of strain,
stiffness decreases and damping increases with strain. This effect appears to become more pronounced
the lower the plasticity of the soil [Vucetic and Dorby, 1991]. Nonlinear, site-specific ground response
analysis requires a numerical approach where the variation of stiffness and damping with shear strain
is described explicitly. Such information is usually determined from cyclic stress-strain tests conducted
on representative soil samples. Ground motions are computed using either an equivalent-linear model
or a more rigorous nonlinear model. In the former, a single solution step is used to predict strain by
selecting stiffness and damping values according to an initially assumed level of strain. Upon computation
of the resulting strains, the stiffness and damping corresponding to the computed strain are checked
against the assumed values and updated. The analysis is repeated until convergence is achieved. The
second formulation involves solution of the equations of motion by nonlinear incremental stress-strain
methods. This approach is potentially more accurate and versatile but often requires more extensive
material characterization, depending on the particular constitutive model selected. The computer code
SHAKE91 is the most widely used program for computing the one-dimensional seismic response of
horizontally layered soil deposits. It uses the equivalent-linear method. A number of other programs are
available that allow for nonlinear dynamic analysis, as well as consideration of excess pore pressures and
the presence of structural elements. Many of the more recent codes, such as PLAXIS, SASSI 2000, and
QUAKE/W, are based on the finite element method.

7.3.3 Design ground motions


The development of design ground motions ideally involves site-specific analysis of the type described
previously. However, a suite of bedrock motions must be considered that corresponds to probable seismic
events ranging from moderate earthquakes with a high probability of occurrence to large earthquakes that
occur only rarely. Seismicity at a particular location can be described in probabilistic terms, accounting for
variations in earthquake magnitude, type, size, and distance. A common requirement is that design ground
motions be based on parameters that have a 10% probability of exceedance in 50 years. Because most seismic
events concentrate energy in a relatively narrow frequency band, it is important that multiple earthquakes
with contrasting spectra be considered so that the computed ground motions cover a broad range of
frequencies. In developing an appropriate set of design ground motions, it is also important to keep in
mind the dynamic characteristics of the structure, acceptable levels of damage, the consequences of failure,
and overall uncertainties [Finn, 2001]. Design ground motions are usually expressed in terms of smooth
response spectra that ignore the small peaks and valleys resulting from any one single seismic record.
Methods of probabilistic seismic hazard analysis are presented in greater detail in Chapter 8.
It is not always possible or desirable to generate design ground motions using site-specific analysis.
For noncritical structures, and as permitted by local ordinances, design ground motion parameters may

© 2003 by CRC Press LLC


0068_C07_fm Page 22 Tuesday, August 13, 2002 1:55 PM

7-22 Earthquake Engineering Handbook

be obtained from applicable building codes. In the United States, building codes are the jurisdiction of
states, which may further delegate adoption and administration of specific provisions to local entities.
Fortunately, most jurisdictions have modeled their provisions on the Unified Building Code (UBC), the
Building Officials Code Administrators (BOCA) National Code, or the Standard Building Code (SBC),
which have now been merged into the International Building Code (IBC) [ICC, 2000]. The seismic
provisions contained in the three earlier model codes share similar formats, particularly in versions after
1994, although specific aspects may vary from location to location. The new International Building Code
was developed in 2000, under the guidance of the International Code Council (ICC), with the intention
of replacing the UBC, BOCA, and SBC codes. Seismic provisions in the IBC code incorporate the
recommendations from the National Earthquake Hazards Reduction Program (NEHRP), developed by
the Building Seismic Safety Council (BSSC). However, the IBC has not yet been widely adopted. At
present, a patchwork of different codes exist across the United States. Nonetheless, the seismic provisions
of the UBC and the NEHRP (BSSC, 1997) are the most common ones in use at present. In both cases,
an allowance is made for site-specific soil conditions in selecting appropriate design accelerations or
spectra. Most countries in seismically active regions have also adopted seismic codes (see Chapter 11).
For example, Japan’s seismic design requirements are contained in its Building Standard Law, most
recently revised in 2000 [Otani, 2000].
In the UBC, earthquake-resistant design of buildings is accomplished by estimating a static design
base shear force that must be accommodated, or by analyzing the dynamic response of the structure
subjected to a smoothed response spectra or a site-specific ground motion time history. The design base
shear force, V, is given by:

1.25 ZIW
V= S (7.10)
RwT 2/3

where
Z = the seismic zone factor
I = the importance factor
Rw = the structural ductility factor
W = the seismic dead load
T = the fundamental period of the structure
S = the soil coefficient
A further description of the parameters in Equation 7.10 can be found in the 1997 edition of the UBC
[ICBO, 1997] and in Chapter 11 of this volume. Of special interest from a geotechnical perspective is
the soil coefficient S, which expresses the site-specific soil conditions. It can be determined from Table 7.1.
This static approach to design should be used with caution because it does not reflect the influence of
soil characteristics on ground motion intensity, frequency, and duration, hence its use is limited by the
UBC to very simple structures. A more comprehensive design can be carried out using one of the

TABLE 7.1 UBC Soil Type Parameter S


Type Description S

S1 A soil profile with either: 1.0


a. A rock-like material characterized by a shear wave velocity greater than 2500 ft/sec or by other suitable
means of classification
OR
b. Medium dense to dense or medium stiff to stiff soil conditions, where the soil depth is less than 200 ft 1.2
S2 A soil profile with predominantly medium dense to dense or medium stiff to stiff soil conditions, where
the soil depth exceeds 200 ft or more
S3 A soil profile containing more than 20 ft of soft to medium stiff clay, but not more than 40 ft of soft clay 1.5
S4 A soil profile containing more than 40 ft of soft clay characterized by a shear wave velocity less than 500 ft/sec 2.0

© 2003 by CRC Press LLC


0068_C07_fm Page 23 Wednesday, July 31, 2002 8:20 AM

Geotechnical and Foundation Aspects 7-23

FIGURE 7.30 Normalized spectral shapes for various types of soil and rock. (From Uniform Building Code. 1994.
Used by permission of the International Conference of Building Officials.)

normalized spectral shapes in Figure 7.30, which were derived from the work of Seed et al. [1976] and
others. One of the spectral curves in Figure 7.30 is chosen for analysis of the structure in question,
depending on the prevailing soil conditions at the site.
The design spectral curve is obtained by multiplying the ordinates in Figure 7.30 by the effective peak
ground acceleration (EPA), defined as EPA = Sa /2.5, where Sa is the average spectral acceleration over
the period interval from 0.1 to 0.5 sec. The EPA can also be assumed as the value of the factor Z in
Equation 7.10, expressed as a function of gravity. When site-specific ground motion histories are to be
used instead of a normalized response spectra, due consideration must be given to selecting a represen-
tative set of ground motions as discussed earlier. Design ground motions in either case must correspond
to a 10% probability of exceedance in a 50-year period. When soil conditions are of type S4, the UBC
requires that a site-specific analysis be conducted for flexible structures whose resonant period is larger
than 0.7 sec.
The UBC site classification system in Table 7.1 is of a qualitative nature. The NEHRP provisions, on
the other hand, allow for a more sound and detailed method of characterizing sites for purposes of
estimating amplification of seismic motions. The main index parameter is the average shear velocity, Vs,
in the upper 30 m (100 ft) of deposit. The motivation for using Vs can be seen by referring to the resonant
period of a soil deposit of thickness H, i.e., T0 = 4H/Vs . Shear velocity is a convenient parameter because
it is inversely proportional to site period and therefore can be used to compare the ground motion
amplification potential of different soil types of similar thickness. The NEHRP site classification scheme
is shown in Table 7.2. Because shear wave velocity measurements may not always be available, alternative
indexes are also provided that correspond approximately to the velocity ranges indicated. The average

standard penetration resistance, N, can be used for cohesionless soils, and the average undrained shear

strength, Su , for clayey soils.
Spectral response shapes can be constructed for any of the soil types in Table 7.2 for various levels of
shaking. Maximum considered earthquake spectral response accelerations at 5% damping are calculated
for short period motions centered at 0.2 sec, SMS, and for longer period motions centered at 1 sec, SM1:

SMS = Fa SS (7.11)

SM1 = Fv S1 (7.12)

© 2003 by CRC Press LLC


0068_C07_fm Page 24 Wednesday, July 31, 2002 8:20 AM

7-24 Earthquake Engineering Handbook

TABLE 7.2 NEHRP and IBC 2000 Site Classifications


Site
Class Description

A Hard rock with measured shear wave velocity: Vs > 5000 ft/sec (1500 m/sec)
– –
B Rock with: 2500 ft/sec Vs ≤ 5000 ft/sec (760 m/sec < Vs ≤ 1500 m/sec)
– –
C Very dense soil and soft rock with 1200 ft/sec < Vs ≤ 2500 ft/sec (360 m/sec < Vs ≤ 760 m/sec) or with
– –
either N > 50 or Su > 2000 psf (100 kPa)
– – –
D Stiff soil with 600 ft/sec ≤ Vs ≤ 1200 ft/sec (180 m/sec ≤ Vs ≤ 360 m/sec) or with either 15 ≤ N ≤ 50 or
– –
1000 psf ≤ Su ≤ 2000 psf (50 kPa ≤ Su ≤ 100 kPa)
– – –
E a. A soil profile with Vs < 600 ft/sec (180 m/sec) or with either N < 15 or Su < 1000 psf (50 kPa)
OR
b. Any soil profile with more than 10 ft (3 m) of soft clay defined as soil with PI > 20, w ≥ 40%, and
Su < 500 psf (25 kPa)
F Soil requiring site-specific evaluations:
1. Soils vulnerable to potential failure or collapse under seismic loading such as liquefiable soils, quick
and highly sensitive clays, collapsible weakly cemented soils
2. Peats and/or highly organic clays (H > 10 ft [3 m] of peat and/or highly organic clay where H = thickness
of soil)
3. Very high plasticity clays (H > 25 ft [8 m] with PI > 75)
4. Very thick soft/medium stiff clays (H > 120 ft [36 m])
Exception: When the soil properties are not known in sufficient detail to determine the Site Class, Site
Class D shall be used. Site Classes E or F need not be assumed unless the authority having jurisdiction
determines that Site Classes E or F could be present at the site or in the event that Site Classes E or F are
established by geotechnical data.

The site coefficients Fa and Fv are obtained from Table 7.3 for various combinations of soil types and
maximum earthquake spectral accelerations. These maximum accelerations are based on a probabilistic
seismic hazard analysis with a 10% probability of exceedance in 50 years. SS and S1 can be obtained from
NEHRP maps prepared by the USGS (http://geohazards.cr.usgs.gov/eq/html/nehrp.html). The maximum
spectral accelerations are scaled for design purposes as follows:

2
SDS = SMS (7.13)
3

2
SD1 = SM1 (7.14)
3

The design response spectrum is prepared as indicated in Figure 7.31 with:

SD1
T0 = 0.2 (7.15)
SDS

SD1
TS = (7.16)
SDS

For periods less than or equal to To , the design spectral acceleration, Sa , is given by:

SDS
Sa = 0.6 T + 0.4SDS (7.17)
To

For periods greater than TS , Sa varies inversely with period:

SD1
Sa = (7.18)
ST

© 2003 by CRC Press LLC


0068_C07_fm Page 25 Tuesday, August 13, 2002 1:55 PM

Geotechnical and Foundation Aspects 7-25

TABLE 7.3A Fa as a Function of Site Class and Short-Period Earthquake Spectral Acceleration
Mapped Maximum Considered Earthquake Spectral Response Acceleration at Short Periods
Site Class Ss ≤ 0.25 Ss = 0.50 Ss = 0.75 Ss = 1.00 Ss ≥ 1.25

A 0.8 0.8 0.8 0.8 0.8


B 1.0 1.0 1.0 1.0 1.0
C 1.2 1.2 1.1 1.0 1.0
D 1.6 1.4 1.2 1.1 1.0
E 2.5 1.7 1.2 0.9 a

F a a a a a

Note: Use straight line interpolation for intermediate values of Ss .


a Site-specific geotechnical investigation and dynamic site response analysis shall be performed.

TABLE 7.3B Fv as a Function of Site Class and 1-Second Earthquake Spectral Acceleration
Mapped Maximum Considered Earthquake Spectral Response Acceleration at 1-Second Periods
Site Class S1 ≤ 0.1 S1 = 0.2 S1 = 0.3 S1 = 0.4 S1 ≥ 0.50

A 0.8 0.8 0.8 0.8 0.8


B 1.0 1.0 1.0 1.0 1.0
C 1.7 1.6 1.5 1.4 1.3
D 2.4 2.0 1.8 1.6 1.5
E 3.5 3.2 2.8 2.4 a

F a a a a a

Note: Use straight line interpolation for intermediate values of S1.


a Site-specific geotechnical investigation and dynamic site response analysis shall be performed.
Spectral Acceleration Sa

SDS

Sa=SD1/T

SD1

To Ts 1.0
Period T (s)
FIGURE 7.31 NEHRP design response spectra. (From BSSC. 1997. NEHRP Recommended Provisions for Seismic
Regulations for New Building and Other Structures, Part 1, Provisions, FEMA 302. Used by permission of the Building
Seismic Safety Council.)

Note that a site-specific procedure for determining ground motion accelerations must be conducted for
all levels of shaking where site conditions of type F are encountered, as well as for very large shaking
intensities at sites of type E (Table 7.3). In general, where a site-specific approach is used, maximum
considered earthquake ground motions must represent a 2% probability of exceedance within a 50-year
period.

© 2003 by CRC Press LLC


0068_C07_fm Page 26 Tuesday, August 13, 2002 1:55 PM

7-26 Earthquake Engineering Handbook

x
z
φ φ1
φ2

ground motion

(A) (B)
FIGURE 7.32 Vertical propagation of horizontal shear wave arriving at a triangular surface wedge. (Modified
from Faccioli, E. 1991. “Seismic Amplification in the Presence of Geologic and Topographic Irregularities,” in Proc.
Second International Conference on Recent Advances in Geotechnical Earthquake Engineering and Soil Dynamics,
St. Louis, MO.)

For the NEHRP static equivalent lateral earth procedure, the base shear force V is calculated as follows:

V = CsW (7.19)

where W includes the dead load and certain other loads, and Cs is the seismic response coefficient that
need not exceed:

SD1
Cs = (7.20)
T (R I )

but should be at least:

Cs = 0.1SD1I (7.21)

For buildings and structures located on sites of types E and F, the minimum value for Cs is:

0.5S1
Cs = (7.22)
RI

In this expression, R is a response modification factor (different from the one used by the UBC), I is an
occupancy factor, and T is the fundamental period of the structure. These parameters are discussed in
greater detail in the NEHRP provisions.
The discussion thus far has focused on sites with a level ground underlain by horizontal soil deposits
of laterally uniform thickness. Evidence suggests that where surface topography is not level or where
large sedimentary basins are encountered ground motions can be complex and significantly different
from what can be expected from conventional one-dimensional attenuation and soil propagation models.
The effect of surface topography can be illustrated with reference to the simple case of upward traveling
SH waves, i.e., shear waves whose motion is perpendicular to the direction of travel, arriving at a
triangular surface wedge of infinite extent (Figure 7.32). Aki [1988] showed that for this case ground
motions can be amplified by a factor of 2π/φ, where the angle φ is measured in degrees. This result can
be used for a first-order estimate of ground motion amplification at ridges and deamplification at troughs
relative to a horizontal topography. Higher levels of ground motions at crests relative to nearby troughs
has indeed been inferred from damage patterns in a number of recent earthquakes [Finn, 1991, 2001].
Sediment deposits in basins are typically thickest near the center and thin out towards the flanks. In
addition to vertically propagating seismic waves, such basins may trap body waves that propagate along

© 2003 by CRC Press LLC


0068_C07_fm Page 27 Wednesday, July 31, 2002 8:20 AM

Geotechnical and Foundation Aspects 7-27

the surface and increase the intensity and duration of the shaking. Records collected in basins often
reveal significant energy at periods above 1 sec. This implies wave lengths much longer than 30 m. Their
amplitude cannot be characterized by the site classification scheme in Table 7.3. Instead, long-period
motion depends on much deeper and larger geologic features, which in many cases are not horizontally
layered. These long-period waves can become trapped in basins that are hundreds of meters thick if they
enter at incidence angles less than critical [Graves, 1993]. The trapped waves propagate across the basin
and can lead to complex interference patterns as they combine with vertically propagating shear waves
and various reflections, particularly near the edges of the basin. Proper modeling in that case requires
that a two- or three-dimensional analysis be conducted [Graves et al., 1998]. The damaging effects from
long-period waves in sedimentary basins were observed during the 1977 Caracas earthquake, the 1985
Michoacan earthquake, the 1994 Northridge earthquake, and the 1995 Kobe earthquake [Somerville,
1998]. Many large cities in the United States, such as Los Angeles, Seattle, Portland, and Tucson, are
located on sedimentary basins and are susceptible to this effect.
Strong ground motions at locations near the earthquake source are strongly influenced by the geometry
of the fault [Somerville, 1998]. Two important effects occur in the near-fault region. When earthquakes
are caused by dipping faults, two sites located on opposite sides but at the same distance from the fault
may experience significantly different motions. Specifically, larger short-period ground motions may
occur on the hanging-wall side of the fault compared to the foot-wall side [Abrahamson and Somerville,
1996]. For earthquakes caused by vertical or dipping faults, the direction of rupture propagation in the
near field can cause substantial differences in the level of shaking for different orientations relative to
the fault’s strike. This has been corroborated by damage patterns in many recent earthquakes but
particularly following the 1995 Kobe earthquake [Finn et al., 1996] and the 1994 Northridge earthquake
[Somerville et al., 1996]. Sites close to the earthquake source typically reveal strike-normal peak velocities
substantially larger than strike-parallel velocities. The effect is most evident at periods longer than about
0.5 sec [Somerville and Graves, 1996]. Obviously, directivity effects must be taken into account in the
design of buildings and other large structures located close to potential earthquake sources.
Site-specific effects are the focus of intense current research efforts. For example, Seed et al. [1997]
and Rodriguez-Marek et al. [1999] have proposed alternate, more intricate soil classification systems that
may aid in the design of ground motions, although they are not yet widely used by practicing engineers.
Near-field site issues are being addressed by seismologists and engineers but have not yet been incorpo-
rated in current building codes. In the future, ground motion maps will likely include the effects described
previously as more data are collected and analyzed, especially now that extensive networks of ground
motion recorders have been installed or are being contemplated worldwide.

7.4 Dynamic Soil Behavior


The behavior of soils under cyclic loading plays a crucial role in determining the extent and distribution
of ground displacements, which in turn are directly related to the potential for causing earthquake
damage. Of importance are the stress-strain and strength response of soil under cyclic conditions. In
general, different levels of induced strain require consideration of different types of material models. If
the focus is on propagation of seismic waves, it is often sufficient to consider small strain, elastic behavior.
On the other hand, if the level of cyclic loading is substantial, it becomes necessary to consider inelastic,
permanent deformations, as well as cyclically induced changes in strength. The cyclic stress-strain and
strength properties of soils can be determined directly or empirically from any of a number of field or
laboratory tests (Table 7.4). The reader is referred to Kramer [1996] for further information.
At relatively low levels of strain, the shear response of a typical soil subjected to cyclic load reversals
of constant amplitude is as shown for any of the stress-strain loops in Figure 7.33A. The hysteretic nature
of the response implies that even at low strains the behavior is nonlinear, albeit reversible, and involves
energy dissipation. For many applications, it is adequate to simplify this type of behavior in terms of
two equivalent parameters that characterize the stiffness and damping aspects of the soil. The equivalent-
linear shear modulus, G, is a function of the cyclic strain amplitude, γc , and is described in terms of a

© 2003 by CRC Press LLC


0068_C07_fm Page 28 Tuesday, August 13, 2002 1:56 PM

7-28 Earthquake Engineering Handbook

TABLE 7.4 Test Methods for Determining Dynamic Properties of Soils


Conditions Test Parameters

Field – Low strain Seismic reflection Vp, thickness of major soil units
Seismic refraction Vs, Vp , thickness of major soil units
Suspension logging V s, V p
Spectral analysis of surface waves VR, wave length, Vs
Seismic cross-hole test Vs, Vp , damping ratio
Seismic down-hole or up-hole test Vs, Vp , thickness of major soil units
Piezocone Vs , thickness of major soil units
Field – High strain Standard penetration test Density, stiffness, strength
Cone penetration test Density, stiffness, strength, soil type, pore pressure
Dilatometer test Stiffness
Lab – Low strain Resonant column Stress-strain, strength
Piezoelectric bender element test Vs
Lab – High strain Cyclic triaxial test Stress-strain, strength, pore pressure
Cyclic direct simple shear test Stress-strain, strength, pore pressure
Cyclic torsional shear test Stress-strain, strength, pore pressure
Shaking table tests Forces and displacements, pore pressure
Centrifuge tests Forces and displacements, pore pressure

Note: Vs = shear wave velocity, Vp = compressional wave velocity, VR = Rayleigh wave phase velocity.

backbone curve (Figure 7.33A and B). Energy dissipation is given in terms of the equivalent damping
ratio, η, which can be expressed as:

1 ED 1 ED
η= = (7.23)
4π E So 2π Gγ s2

where ESo is the strain energy stored in the soil at a cyclic strain of γc , and ED is the energy dissipated
within the hysteretic loop. The equivalent damping ratio is also a function of strain level, γc , as indicated
in Figure 7.33C. The maximum value of the shear modulus, Gmax , represents the tangent stiffness at very
low strain (less than about 10 –4 %). It is generally difficult to measure from the stress-strain response of
laboratory specimens, which typically involve relatively large levels of strain but instead is obtained from
shear wave velocity measurements conducted either in the field or in laboratory samples:

Gmax = ρVs2 (7.24)

where ρ is the density of the soil. Alternatively, Gmax can be obtained empirically from any of a number
of in situ field tests, which are conveniently summarized in Kramer [1996]. Some guidelines are provided
in the section on liquefaction settlement.
The equivalent-linear model is appealing because it incorporates material nonlinearity, yet it only
requires two relatively simple material constants. It is easily implemented and commonly used for ground
response analysis. Nonetheless, it is important to realize that the model represents an approximation to
soil behavior. It assumes that no permanent displacements accumulate with multiple cycles of loading
and that no failure occurs. True cyclic soil behavior for load reversals of constant amplitude is closer to
that shown in Figure 7.34. It involves open hysteretic loops with irreversible energy dissipation and
permanent strain accumulation. This type of behavior can be accounted for but it requires a point-wise
description of the backbone curve and a set of rules for unloading and reloading behavior [Kramer,
1996]. A hierarchy of nonlinear cyclic models can be considered, ranging from simple Masing-type
models with simple loading-unloading rules, to more sophisticated models that can account for changes
in the shape of the backbone curve due to volume changes, i.e., hardening or softening, and for mech-
anisms of pore pressure generation that lead to changes in effective stress during cyclic loading. Although
quite powerful, sophisticated nonlinear cyclic models are typically limited in terms of soil types and
loading conditions for which they provide reasonable predictions.

© 2003 by CRC Press LLC


0068_C07_fm Page 29 Wednesday, July 31, 2002 8:20 AM

Geotechnical and Foundation Aspects 7-29

(A) τ

Gmax G

τc Backbone curve

ESo γ
ED γc

(B) (C)

G
Gmax η

log γ log γc

FIGURE 7.33 (A) Typical response of soil subjected to cyclic load reversals; (B) modulus reduction curve; and
(C) damping curve.

FIGURE 7.34 Approximate true behavior of soil subjected to load reversals of constant amplitude.

© 2003 by CRC Press LLC


0068_C07_fm Page 30 Wednesday, July 31, 2002 8:20 AM

7-30 Earthquake Engineering Handbook

1.2 Single amplitude failure strain:


εf >5%
1.0

Cyclic strength ratio, τcyc/Su


εf ≤3%

0.8

0.6

0.4

0.2

0
1 3 10 100 300 1,000 3,000 10,000

Number of cycles to failure, N

FIGURE 7.35 Cyclic strength ratio variation with number of cycles to reach failure strain for different soils. (From
Lee, K.L. and J.A. Focht. 1976. “Strength of Clay Subjected to Cyclic Loading,” Mar. Geotechnol., 1(3), 165–188. Used
by permission of Taylor and Francis, Inc.)

The most thorough approach to modeling the cyclic behavior of soils involves the use of general elasto-
plastic constitutive models that account for three-dimensional stress and strain components, the effect
of initial stress conditions, general stress paths, hardening or softening behavior, and generation and
dissipation of excess pore pressures under monotonic and cyclic loading conditions. They are often
framed in terms of critical state concepts. Even though they are potentially very accurate and general in
nature, they have not found widespread use in standard practice because they typically require a large
number of material parameters, many of which are difficult to determine. However, comprehensive soil
models can provide a compatible link between pre-failure and failure deformation mechanics.
It is useful to consider the strength of fine-grained cohesive soils separately from that of noncohesive
granular soils that may be susceptible to liquefaction. Failure associated with the phenomenon of lique-
faction is discussed in the next section. For all practical purposes, cyclic loading due to ground shaking
can be assumed to occur under undrained conditions in most soils, except perhaps in coarse gravels.
Failure is usually defined in terms of a critical level of deformation, beyond which the ability of a particular
structure to provide its intended service is compromised. This may involve either cyclic or unidirectional
displacements that exceed acceptable thresholds. The maximum shear stress associated with this level of
straining is referred to as the cyclic strength. It is a function of the average shear stress, as well as the
amount of shear stress induced in each cycle of loading. Cyclic shear strength is usually expressed in
terms of the cyclic strength ratio, τcyc /Su , where Su is the undrained shear strength under monotonic
loading conditions. If the average shear stress is zero, no unidirectional shearing occurs and the strength
ratio associated with a particular level of failure shear strain (typically taken in the 3 to 5% range)
decreases with the number of load cycles (Figure 7.35). Note that at cyclic stress levels below a given
value, the failure shear strain is never reached and the soil is said to remain stable. As is evident from
Figure 7.35, this cutoff stress ratio varies with soil type. It generally increases with plasticity.
If the average shear stress is not zero, as is the case along potential failure surfaces associated with
slopes, retaining structures, and most types of foundations, the cyclic shear strain and the average shear
strain will depend on both the average level of shear stress and the cyclic shear stress. For example, a
potential failure plane with an average static shear stress level that is close to the maximum shear stress
that can be mobilized requires only a small amount of cyclic loading to reach unacceptable levels of shear
strain, whereas one that is subjected to a low level of average shear stress will require more intensive
cyclic loading to reach the same level of failure shear strain.
After earthquake shaking ceases, it is necessary to make sure that critical structures remain stable over
the long term. Any changes in available shear strength must be evaluated. The large-strain, residual

© 2003 by CRC Press LLC


0068_C07_fm Page 31 Wednesday, July 31, 2002 8:20 AM

Geotechnical and Foundation Aspects 7-31

100

80 Safe

Liquid Limit (%)


60

40
Liquid Limit = 35%
0.9 x Liquid Limit
20
Test

0
0 20 40 60 80 100

Water Content (%)

FIGURE 7.36 Modified Chinese Criteria for liquefaction assessment of fine-grained soils. (Modified from Finn,
W.D.L., R.H. Ledbetter et al. 1994. Liquefaction in Silty Soils: Design and Analysis, Geotechnical Publication No. 44,
American Society of Civil Engineers, pp. 51–76.)

strength of saturated soil depends on void ratio and fabric. If shaking does not induce changes in either
of these two, undrained strength before and after shaking is essentially the same. Although stiffness may
be affected by the development of excess pore pressures, volume changes that follow the shaking deter-
mine changes in strength; that is, of course, if no disturbance of the soil structure occurs. Soil destruc-
turing is of little importance in most insensitive soils strained to low or modest strain levels but it becomes
increasingly more important as cyclic strain amplitude increases and it is certainly a major concern in
sensitive soils [Thiers and Seed, 1978].

7.5 Liquefaction

7.5.1 General Concepts


The phenomenon of liquefaction is generally associated with cohesionless soils. It results from seismic
shaking that is of a sufficient intensity and duration. It occurs most commonly in loose, saturated,
granular soils that are uniformly graded and that contain few fines. Although sands are especially
susceptible, liquefaction is also known to develop in some silts and gravels. Two necessary conditions for
liquefaction to occur are the presence of soils of sufficiently low density that will tend to undergo volume
reduction upon shaking, and a state of full or near full saturation. Under these conditions, cohesionless
soils will tend to densify when subjected to cyclic shear stresses from ground vibrations but will be
temporarily prevented from doing so at depth due to restricted drainage. As a result, excess pore pressures
accumulate, effective stresses decrease, and soils lose strength and may become liquefied [Seed and Idriss,
1982]. Because the capacity of soils to bear foundation loads is directly related to their strength, lique-
faction poses a serious hazard to constructed structures and must be assessed in seismic areas where
susceptible deposits exist. It should be noted that liquefaction herein refers to loss of strength and stiffness
due to the cyclic pore pressure-generation mechanism just described, and not due to strength loss and
stiffness changes that occur in sensitive soils upon monotonic shearing or remolding. This latter mech-
anism presents a danger in its own right but is not normally seismically related.
Not all granular soils are prone to liquefaction. As a general rule of thumb, cohesionless deposits with
depth-corrected standard penetration values (N1)60 > 30, depth-corrected normalized cone penetration
resistance values qc1N > 175, or stress-corrected shear wave velocity VS1 > 230 m/sec (755 ft/sec) are
considered of sufficient density to pose little risk of liquefying. The potential for fine-grained soils to
liquefy can be evaluated with reference to the Modified Chinese Criteria [Finn et al., 1994] shown in
Figure 7.36. According to these criteria, soils may liquefy if the clay fraction is less than 15% (using the

© 2003 by CRC Press LLC


0068_C07_fm Page 32 Tuesday, August 13, 2002 1:56 PM

7-32 Earthquake Engineering Handbook

Liquid Limit < 32 Liquid Limit ≥ 32


Clay content < 10% Susceptible May be susceptible
(conduct additional testing)

Clay content ≥ 10% May be susceptible Not susceptible


(conduct additional testing)
FIGURE 7.37 Liquefaction criteria for fine-grained soils. (Modified from Andrews, D.C.A. and G.R. Martin. 2000.
“Criteria for Liquefaction of Silty Soils,” in Proc. 12th World Conference on Earthquake Engineering, Auckland,
New Zealand.)

Chinese definition of clay size as less than 0.005 mm), the liquid limit is less than 35%, and the water
content is larger than 0.9 times the liquid limit. These criteria are somewhat controversial and they do
not represent a consensus among practicing engineers [NCEER, 1997]. Establishing more precise and
reliable measures for identifying which finer-grained soils are potentially susceptible to liquefaction is
an area of ongoing research [Seed et al., 2001]. For example, Andrews and Martin [2000] have reevaluated
a large number of field liquefaction case histories and have proposed an adaptation of the Modified
Chinese Criteria for U.S. use (Figure 7.37).
Evaluation of the liquefaction resistance of soil deposits has evolved over the years into a commonly
accepted, simplified empirical procedure, originally developed by Seed and Idriss [1971]. This simplified
procedure represents the standard of practice in North America and in many other countries. It is
reviewed periodically by groups of experts who make recommendations for changes according to data
collected from new earthquakes and new developments in liquefaction hazard assessment. The latest
consensus statement is contained in a 1996 NCEER workshop report [NCEER, 1997; Youd and Idriss,
2001], upon which much of the following discussion is based.
The potential for liquefaction is assessed with the aid of liquefaction charts, which are based on
observations of whether liquefaction did or did not occur at specific sites during numerous past earth-
quakes. These charts can be used to determine what combinations of shaking intensity and soil resistance
are likely to result in liquefaction. Cyclic resistance ratio (CRR) curves represent limiting conditions
that determine whether liquefaction will occur. The intensity of earthquake shaking is expressed in terms
of the average effective cyclic stress ratio (CSR). Seed and Idriss [1971] proposed the following expression
for CSR:

τ av a σ
CSR = = 0.65 max vo rd (7.25)
σ ′vo g σ ′vo

where τav is the equivalent shear stress amplitude, amax is the peak horizontal acceleration at ground
surface generated by the earthquake, g is the acceleration of gravity, σvo and σ′vo are the total and effective
vertical overburden stresses, respectively, and rd is a nonlinear stress reduction coefficient that varies with
depth and may be approximated as follows:

rd = 1.000 − 0.00765 z z ≤ 9.15 m (7.26a)

rd = 1.174 − 0.02670 z 9.15m < z ≤ 23m (7.26b)

rd = 0.744 − 0.00800 z 23m < z ≤ 30m (7.26c)

rd = 0.500 z > 30m (7.26d)

© 2003 by CRC Press LLC


0068_C07_fm Page 33 Wednesday, July 31, 2002 8:20 AM

Geotechnical and Foundation Aspects 7-33

FIGURE 7.38 Liquefaction resistance based on SPT. (From Seed, H.B., K. Tokimatsu et al. 1985. “The Influence of STP
Procedures in Soil Liquefaction Resistance Evaluations,” J. Geotech. Eng. ASCE, 111(12), 1425–1445. With permission.)

These expressions represent mean values that are suitable for routine engineering practice. However,
the variance in rd at any given depth is considerable, and the reliability of these expressions has not
been sufficiently verified at depths below about 15 m. In fact, Cetin and Seed [2000] have suggested
that the above equations overestimate the true value of rd at depths between 3 and 10 m. They propose
a new empirical method for determining rd . However, their approach has not yet been widely adopted
in practice. In any case, the best method of estimating the in situ CSR is to conduct a site-specific
analysis for a particular earthquake of interest, if feasible. In the simplified procedure, soil resistance
to liquefaction is evaluated using any of a number of in situ tests, including the standard penetration
test (SPT), the cone penetration test (CPT), shear wave velocity measurements (Vs), and the Becker
penetration test.
7.5.1.1 SPT Liquefaction Assessment Chart
Liquefaction resistance based on SPT results can be determined with reference to Figure 7.38. This figure
was originally developed by Seed and coworkers [1985] and is continually updated with observations

© 2003 by CRC Press LLC


0068_C07_fm Page 34 Tuesday, August 13, 2002 1:56 PM

7-34 Earthquake Engineering Handbook

TABLE 7.5 SPT Corrections


Correction Factor Variable Term Value

Energy ratio a Donut hammer CE 0.5–1.0


Safety hammer 0.7–1.2
Automatic-trip donut-type hammer 0.9–1.3
Borehole diameter 65–115 mm CB 1.0
150 mm 1.05
200 mm 1.15
Rod length 3–4 m CR 0.75
4–6 m 0.85
6–10 m 0.95
10–30 m 1.0
>30 m <1.0
Sampling method Standard sampler CS 1.0
Sampler without liner 1.1–1.3
a The preferred method is to measure the energy ratio repeatedly at each site.

following new earthquakes. Combinations of corrected blow count (N1)60 and CSR that plot above the
appropriate CRR curve indicate a high probability of liquefaction, whereas those plotting below suggest
a low probability of liquefaction. Individual CRR curves are shown for various amounts of fines. The
CRR curve for a fines content of less than 5% is referred to as the ”simplified base curve.” The SPT chart
is for an earthquake of magnitude 7.5. Corrections for earthquakes of different magnitude, as well as for
high overburden pressures and sloping ground surface, are discussed later.
The raw measured penetration number, Nm , is corrected to a standard set of conditions so that field
values can be compared in a meaningful way. Nm is adjusted to correspond to 60% of the maximum
potential free fall energy of a U.S. safety hammer operated by rope and pulley. A correction factor, CE ,
is applied to the penetration number for other combinations of hammers and hammer release modes.
Although values are suggested for use below, it is preferable to directly measure the energy ratio at each
site where the SPT is used. Adherence to ASTM D-1686 procedures during field testing should yield
consistent energy ratios. In granular soils, N is also affected by the effective overburden pressure σ ′vo ,
therefore it is corrected to a standard stress of 95.6 kN/m2 (1 ton/ft2) through the use of the factor CN.
Corrections to N values are also made for borehole diameter (CB), rod length (CR ), and sampling method
(CS):

(N1 )60 = CN N mCECBCRCS (7.27)

(N1)60 represents the corrected standard penetration value to be used with the SPT liquefaction chart.
Values of CE, CB , CR, and CS recommended by the NCEER committee [NCEER, 1997] are listed in
Table 7.5. A number of different expressions exist for the overburden correction factor CN. One of the
most common ones is given by Liao and Whitman [1986]:

Pa
CN = (7.28)
σ ′vo

where Pa equals 1 U.S. ton/ft2, or 96 kPa, and is given in the same units as the overburden stress σ′vo .
Values of CN should not exceed about 2.0 for very shallow deposits.
7.5.1.2 CPT Liquefaction Assessment Chart
Resistance of soil deposits to liquefaction can also be determined from tip resistance measured during a
cone penetration test (CPT). As opposed to the SPT, the CPT provides continuous profiles of penetration
resistance. This represents a distinct advantage for interpreting the stratigraphy of soil deposits because
the location of soil boundaries and the presence of thin soil layers can be detected much more accurately.

© 2003 by CRC Press LLC


0068_C07_fm Page 35 Tuesday, August 13, 2002 1:56 PM

Geotechnical and Foundation Aspects 7-35

FIGURE 7.39 Liquefaction criteria based on CPT data. (From Robertson, P.K. and C.E. Wride. 1997. “Cyclic
Liquefaction and its Evaluation Based on the SPT and CPT,” in Proc. NCEER Workshop on Evaluation of Liquefaction
Resistance of Soils, Salt Lake City, NV. Used by permission of the Multidisciplinary Center for Earthquake Engineering
Research, Buffalo, NY.)

CPT results are generally more consistent and repeatable than those from any of the other tests used for
liquefaction assessment. In addition, the CPT test is conducted under much lower rates of penetration
than the SPT, therefore it is inherently more suitable for providing empirical estimates of soil behavior
parameters, including those associated with seismic response. However, CPT criteria for evaluating
liquefaction resistance are not yet as widely accepted as those associated with the SPT. The reason is that
the database upon which the liquefaction criteria in the corresponding CPT chart (Figure 7.39) are based
is not nearly as extensive as it is for the SPT chart. There is no doubt that as new field performance data
become available and the correlations in Figure 7.39 are revised, the CPT approach will find a wider
acceptance in engineering practice
The CPT liquefaction assessment chart in Figure 7.39, developed by Robertson and Wride [1997], is
used to determine the CRR for a magnitude 7.5 earthquake and for clean sands with less than 5% of
fines. The dashed curves show approximate cyclic shear strain potential, γe , to emphasize that shear strain
and the potential for ground deformation increase as the resistance to penetration decreases. The raw
cone penetration resistance, qc , is corrected for overburden stress and normalized to qc1N , as follows:

qc
qc1N = CQ (7.29)
Pa

n
 P 
CQ =  a  (7.30)
 σ ′vo 

where Pa equals 1 U.S. ton/ft2, or 96 kPa, and is given in the same units as the overburden stress σ ′vo . A
maximum value of 2.0 is applied to the normalizing factor CQ at shallow depths. The exponent n has a
value that depends on soil type and ranges from about 0.5 for clean sands to about 1.0 for clayey-type
soils. More details on this parameter and the influence of soil type on qc1N are given in Robertson and

© 2003 by CRC Press LLC


0068_C07_fm Page 36 Wednesday, July 31, 2002 8:20 AM

7-36 Earthquake Engineering Handbook

Wride [1997]. An additional correction to cone penetration resistance for the presence of thin layers of
sand or stiff clay, bounded on either side by softer soil, is discussed by Robertson and Fear [1995].
7.5.1.3 Shear Wave Velocity Liquefaction Assessment Chart
Shear wave velocity Vs can be used as an index to liquefaction resistance because both Vs and CRR reflect,
albeit in different proportions, a soil deposit’s void ratio, effective confining stress, stress history, and
geologic age [Youd and Idriss, 2001]. Although shear velocity can be determined with great accuracy in
the field by various downhole, crosshole, seismic CPT, and surface wave spectral analysis techniques, it
reflects small strain behavior, whereas liquefaction involves large strain phenomena. Shear wave velocity
measurements should be conducted in conjunction with a limited number of borings when assessing
liquefaction resistance so that soil intervals that would appear to classify as liquefiable by the Vs criteria
do not include nonliquefiable, soft, clay-rich soils [NCEER, 1997]. Similarly, weakly cemented soils may
be susceptible to liquefaction even though they classify as nonliquefiable based on measured values of Vs .
Shear wave velocity is normalized in a similar manner to the SPT and CPT procedures described above:

0.25
 P 
VS1 = VS  a  (7.31)
 σ ′vo 

where again Pa equals 1 U.S. ton/ft2, or 96 kPa, and is given in the same units as the overburden stress
σ′vo . The liquefaction chart in Figure 7.40 should only be used for Holocene-age uncemented soils. A
dashed CRR curve is shown for values of Vs less than 100 m/sec to indicate that essentially no field
performance data exists at lower velocities to substantiate the indicated trend.
7.5.1.4 Becker Penetration Test
Gravelly soils can be susceptible to liquefaction. However, SPT, CPT, and Vs measurements are unreliable
in these types of soils. The reason is that as the size of individual soil particles approaches that of the
penetrometer, resistance to penetration is tied to the properties and arrangement of individual particles
in the zone of deformation. This can increase the resistance to penetration in unpredictable ways. The
obvious solution appears to be the use of larger penetrometers. The Becker Penetrometer Test (BPT)
constitutes a type of test that makes use of this concept. It consists of a 168-mm-wide double-walled
casing that is driven into the ground by a diesel pile hammer. The Becker penetration resistance is defined
as the number of blows required to advance the casing a distance of 300 mm. Very few tests have been
conducted at sites where liquefaction has been observed to occur and no liquefaction charts equivalent
to those in Figures 7.38 and 7.39 have been developed. Instead, the BPT has been correlated to the SPT
at a number of locations where both types of tests have been conducted side by side [Harder and Seed,
1986]. Using the correlations in Figure 7.41, the BPT value is converted to an equivalent SPT blow count,
which in turn is used to estimate liquefaction resistance.
There are a number of concerns about the BPT that have not been addressed adequately. Due to the
high driving energy that is possible with this device, the test is appealing for use in very deep deposits.
However, friction along the casing becomes difficult to evaluate properly in deposits deeper than about
30 m. In this case, the correlation of Harder and Seed [1986] in Figure 7.41, which incorporates friction
intrinsically, may not be appropriate [Sy and Campanella, 1994]. Also, this correlation should not be
used for sites with thick, dense deposits overlying loose sands or gravels because it has not been verified
for these conditions [NCEER, 1997].

7.5.2 Magnitude Scaling Factors and Other Corrections


The CRR curves in the SPT, CPT, and Vs charts correspond to an earthquake of magnitude 7.5. Seed and
Idriss [1982] suggested the use of magnitude scaling factors (MSF) for earthquakes of magnitude other
than 7.5. These factors are used to shift the CRR base curve vertically according to:

CRRcor = CRR 7.5 × MSF (7.32)

© 2003 by CRC Press LLC


0068_C07_fm Page 37 Tuesday, August 13, 2002 1:57 PM

Geotechnical and Foundation Aspects 7-37

FIGURE 7.40 Liquefaction criteria based on shear wave velocity for Holocene uncemented sands. (From Andrus,
R.D. and K.H. Stokoe. 1997. “Liquefaction Resistance Based on Shear Wave Velocity,” in Proc. NCEER Workshop on
Evaluation of Liquefaction Resistance of Soils, Salt Lake City, NV. Used by permission of the Multidisciplinary Center
for Earthquake Engineering Research, Buffalo, NY.)

The range of MSF values recommended by the NCEER committee is shown in Figure 7.42.
The chart-based empirical approach to liquefaction assessment has become widely used in engineering
practice. It was originally intended for relatively shallow and level deposits. In order to expand the range
of conditions for which the approach can be applied, correction factors were developed by Seed [1983]
for large static overburden stresses and for large shear stresses on potential failure planes, specifically for
use in conjunction with liquefaction assessment of embankment dams. When applicable, these factors
are used as follows:

CRRcor = CRR 7.5× MSF × K σ × K α (7.33)

Very few liquefaction observations have been made where large initial overburden or shear stresses played
a role. Therefore, these factors have been developed primarily from the results of extensive laboratory
testing.
The liquefaction resistance of a soil increases with increasing confining pressure. Cyclic triaxial com-
pression tests show that the resistance to liquefaction, as measured by the CSR, decreases nonlinearly
with increasing confining pressure. To account for this effect at overburden pressures larger than 100 kPa
(depths larger than about 10 to 15 m), Seed [1983] recommended the use of the factor Kσ. The NCEER
committee [1997] reviewed the Seed factors and recommended that the values shown in Figure 7.43 be
used instead. They also reviewed data from a number of studies that considered the effect of soil type
and density but no consensus was reached on what to recommend for engineering practice. The NCEER

© 2003 by CRC Press LLC


0068_C07_fm Page 38 Tuesday, August 13, 2002 1:57 PM

7-38 Earthquake Engineering Handbook

FIGURE 7.41 Correlation between BPT and SPT blow counts. (From Harder, L.F. and H.B. Seed. 1986. “Determi-
nation of Penetration Resistance for Coarse-Grained Soils Using the Becker Hammer Drill,” Earthquake Engineering
Research Center, University of California, Berkeley. With permission.)

FIGURE 7.42 Range of MSF values. (From Youd, T.L. and S.K. Noble. 1997. “Magnitude Scaling Factors,” in Proc.
NCEER Workshop on Evaluation of Liquefaction Resistance of Soils, Salt Lake City, NV. Used by permission of the
Multidisciplinary Center for Earthquake Engineering Research, Buffalo, NY.)

committee is considering three correction curves, with the curve shown in Figure 7.43 corresponding to
a medium-dense sand [Finn, 2001]. In any case, caution must be exercised when using these factors at
depths exceeding 15 m because of the lack of field data to substantiate the reported values.
Where a sloping ground surface exists, potential failure surfaces are subjected to gravitational shear
stresses prior to earthquake shaking. The effect of such an initial static shear stress, τst , on subsequent
liquefaction resistance can be considered by normalizing such a stress by the effective vertical stress, i.e.,

© 2003 by CRC Press LLC


0068_C07_fm Page 39 Tuesday, August 13, 2002 1:57 PM

Geotechnical and Foundation Aspects 7-39

FIGURE 7.43 Correction for effective confining pressure for clean and silty sands and gravels. (From Harder, L.F.
and R. Boulanger. 1997. “Application of Ksigma and Kalpha Correction Factors,” in Proc. NCEER Workshop on Evaluation
of Liquefaction Resistance of Soils, Salt Lake City, NV. Used by permission of the Multidisciplinary Center for
Earthquake Engineering Research, Buffalo, NY.)

2.0

σ'vo ≤ 300 kPa

1.5
Dr≈55-70%


1.0
Dr≈45-50%

0.5 Dr≈35%

0.0
0.0 0.1 0.2 0.3 0.4
τ
α= s
σ'vo

FIGURE 7.44 Effect of initial static shear stress on liquefaction potential. (From Harder, L.F. and R. Boulanger.
1997. “Application of Ksigma and Kalpha Correction Factors,” in Proc. NCEER Workshop on Evaluation of Liquefaction
Resistance of Soils, Salt Lake City, NV. Used by permission of the Multidisciplinary Center for Earthquake Engineering
Research, Buffalo, NY. With permission.)

α = τst /σ′vo . This alpha ratio is zero for a level ground surface. In general, its has been found that the
presence of a static shear stress on the potential failure surface tends to increase the cyclic resistance, in
terms of the CSR necessary to trigger liquefaction, as long as the soil is relatively dense and the confining
pressure is low. On the other hand, loose soils and some soils under high confining pressure have a lower
liquefaction resistance when an initial static shear stress is present [Harder and Boulanger, 1997; NCEER
1997]. Harder and Boulanger have recommended tentative α vs. Kα curves (Figure 7.44), which, however,
were not endorsed by the NCEER committee for use in practice due to their uncertainty. Liquefaction
analysis for sloping ground conditions is a complex problem that requires further research.

© 2003 by CRC Press LLC


0068_C07_fm Page 40 Wednesday, July 31, 2002 8:20 AM

7-40 Earthquake Engineering Handbook

It has been observed that liquefaction resistance of soil deposits increases with age [Youd and Hoose,
1977; Youd and Perkins, 1978; Seed, 1979b]. Young Holocene sediments are more prone to liquefaction
than older Holocene sediments. Pleistocene and earlier deposits are even more resistant. However, this
trend has not been quantified due to a lack of data and an incomplete understanding of the effects of
age on liquefaction resistance. The liquefaction assessment procedures described above should only be
used for Holocene deposits that are not more than a few thousand years old. In some cases, it may be
possible to balance the effect of a decreasing value of Kσ with depth with a parallel increase in liquefaction
resistance that could be expected due to increasing age with depth. Such an approach should only be
used with great care and certainly would only be reasonable for natural deposits where age effects increase
uniformly with depth [Youd and Idriss, 2001].

7.5.3 Liquefaction Settlement


Loose sand deposits that are prone to liquefaction during an earthquake are also susceptible to settlement,
which has proven to be very damaging to structures, pavements, and lifelines. Recent earthquakes in
Japan, Turkey, Taiwan, and elsewhere have resulted in surface settlements of tens of centimeters at some
locations. However, estimating settlement resulting from seismic shaking is a difficult proposition and
errors of 50% or more can be expected when making predictions [Kramer, 1996]. Although detailed
ground response analysis can be conducted to estimate settlements, and this may be necessary for dams
or other instances where the ground surface is nonlevel, a simplified procedure may be sufficient in many
cases [Tokimatsu and Seed, 1987].
Liquefaction settlement in dry sands is a function of the density of the soil, the number of strain cycles
and the magnitude of the cyclic shear strain induced by seismic shaking [Silver and Seed, 1971]. An
effective shear strain, γeff , can be calculated from the average cyclic shear stress, τav , as follows:

τ av
γ eff = (7.34)
 Geff 
Gmax  
 Gmax 

where G max is the small-strain shear modulus and G eff the effective shear modulus at the induced strain
level. The average cyclic shear stress is given by Equation 7.25; therefore, the above expressions can be
rewritten as:

G  a σ
γ eff  eff  = 0.65 max vo rd (7.35)
 Gmax  g Gmax

G max can be determined from shear wave velocity measurements or other suitable small-strain laboratory
or field procedures. It can also be approximated by the relationship given by Seed and Idriss [1970]:

Gmax = 1, 000 × ( K 2 )max × (σ m′ )


1/3
(7.36)

where σ′m is the mean principal stress in units of psf. The parameter (K2)max is a function of void ratio
or relative density (Table 7.6) and can also be obtained from corrected SPT values:

(K 2 )max = 20[(N1 )60 ]


1/3
(7.37)

Equations (7.36) and (7.37), along with the appropriate maximum acceleration, amax, are evaluated to
determine the term on the left-hand side of Equation (7.35).
This quantity is used in Figure 7.45 to determine the effective shear strain γeff , which in turn is used
with Figure 7.46 to obtain the associated volumetric strain. A soil profile is divided into an appropriate
number of layers of approximately uniform relative density, or corrected SPT number, and the volumetric
strain is computed using the above approach. Figure 7.46 corresponds to an earthquake of magnitude 7.5

© 2003 by CRC Press LLC


0068_C07_fm Page 41 Wednesday, July 31, 2002 8:20 AM

Geotechnical and Foundation Aspects 7-41

TABLE 7.6 (K2) max Values for Estimating Maximum Shear Modulus
DR (%) (K2) max Void ratio, e (K2) max

30 34 0.4 70
40 40 0.5 60
45 43 0.6 51
60 52 0.7 44
75 59 0.8 39
90 70 0.9 34

Source: Modified from Seed, H.B. and I.M. Idriss. 1970. “Soil Moduli and Damping
Factors for Dynamic Response Analysis,” Earthquake Engineering Research Center, Uni-
versity of California, Berkeley.

10−2

σ′m = 0.1 tsf 0.2


0.5
12

10−3
Shear Strain, γeff

10−4

10−5
10−5 10−4 10−3
γeff (Geff/Gmax)

FIGURE 7.45 Induced strain in sand deposits. (From Tokimatsu, K. and H.B. Seed. 1987. “Evaluation of Settle-
ments in Sand Due to Earthquake Shaking,” J. Geotech. Eng. ASCE, 113(8), 861–878. Used by permission of the
American Society of Civil Engineers.)

with approximately 15 representative cycles. For earthquakes with magnitudes other than 7.5, the volu-
metric strain can be obtained from the volumetric strain scaling factor rm :

rm =
(ε c )M (7.38)
(εc )M =7.5
Values for r m are listed in Table 7.7. Since the relationships in Figure 7.46 are based on unidirectional
simple shear conditions, whereas actual earthquakes subject soil deposits to multidirectional shaking, it
is necessary to double the volumetric strains obtained from Figure 7.46. Settlement is calculated by
multiplying the volumetric strain by the layer thickness. The total settlement then is the sum of the
settlements for each of the layers. Finally, the total settlement is doubled to account for the effects of
multidirectional shaking.

© 2003 by CRC Press LLC


0068_C07_fm Page 42 Wednesday, July 31, 2002 8:20 AM

7-42 Earthquake Engineering Handbook

Cyclic Shear Strain, γxy − percent


−3
10 10−2 10−1 1
10−3
N1 ≈ 40 15 Cycles
≈ 30

Volumetric Strain Due to Compaction, εc − percent


≈ 20
≈ 15
10−2 ≈ 10
≈5

10−1

10

FIGURE 7.46 Relationship between volumetric strain, shear strain, and penetration resistance for dry sands. (From
Tokimatsu, K. and H.B. Seed. 1987. “Evaluation of Settlements in Sand Due to Earthquake Shaking,” J. Geotech. Eng.
ASCE, 113(8), 861–878. Used by permission of the American Society of Civil Engineers.)

For the case of saturated sands, Tokimatsu and Seed [1987] provide the chart in Figure 7.47 to estimate
the corresponding volumetric strain of each soil layer. Again, this strain is adjusted for earthquakes of
magnitude other than 7.5 by using the factors in Table 7.7 and the settlements are calculated as described
for the case of dry sand.

7.5.4 Shear Strength of Liquefied Soil


Where the ground surface is inclined, as in the case of natural or embankment slopes, post-liquefaction
lateral displacements are a potential concern. The amount of lateral deformation following a period of
shaking is a function of the difference between the post-liquefaction steady-state shear strength, Ssu , and
the shear stress necessary to reestablish equilibrium. The larger this difference, the larger the expected
liquefaction flow. The residual shear strength corresponds to a limiting state of constant shear stress, constant
confining pressure, and constant volume [Castro and Poulos, 1977]. It can be determined through laboratory
procedures or field correlations. It is a function primarily of the volume change tendency of the soil.
The laboratory procedure consists of consolidating a representative specimen to its in situ stress state,
as close as possible, and shearing it under undrained conditions along an appropriate stress path. The
resulting steady-state shear strength then must be corrected for disturbance due to sampling and han-
dling. Poulos et al. [1985] have suggested a rational procedure for carrying out such a correction that
involves additional testing of reconstituted specimens. Conversely, Seed [1986] analyzed a number of
flow slides and back-calculated the post-liquefaction shear strength in each case. He suggested a rela-
tionship between shear strength and SPT penetration number (Figure 7.48).
Seed and Harder [1990] and Stark and Mesri [1992] have further proposed that corrections be applied
to the penetration number for soils with more than 10% fines to obtain an equivalent clean sand SPT
value for use in Figure 7.48:

(N1 )60−cs = (N1 )60 + N corr (7.39)

where the correction factor N corr is obtained from Table 7.8.

© 2003 by CRC Press LLC


0068_C07_fm Page 43 Tuesday, August 13, 2002 1:57 PM

Geotechnical and Foundation Aspects 7-43

0.6

Volumetric strain (%)


10 5 4 3 2 1 0.5
0.5

0.2
0.4

0.1
τav
σ'vo 0.3

0.2

0.1

0.0
0 10 20 30 40 50

(N1)60

FIGURE 7.47 Relationship between cyclic stress ratio and volumetric strain for saturated clean sands. (From Toki-
matsu, K. and H.B. Seed. 1987. “Evaluation of Settlements in Sand Due to Earthquake Shaking,” J. Geotech. Eng.
ASCE, 113(8), 861–878. Used by permission of the American Society of Civil Engineers.)

TABLE 7.7 Scaling Factors for Equation 7.38


Earthquake Volumetric Strain
Magnitude (M) Scaling Factor (rm)

5.25 0.4
6 0.6
6.75 0.85
7.5 1.0
8.5 1.25

Source: Adapted from Tokimatsu, K. and H.B. Seed.


1987. “Evaluation of Settlements in Sand Due to Earth-
quake Shaking,” J. Geotech. Eng. ASCE, 113(8), 861–878.

7.6 Seismic Analysis of Slopes and Dams


Seismic shaking can induce destabilizing dynamic forces in earth embankments, natural and constructed
slopes, and retained fills. The result may be unacceptable levels of deformation and even outright failure.
A series of analysis methods exist to evaluate the effects of earthquake motions, ranging from a simple
pseudostatic approach to comprehensive stress-strain dynamic finite element methods that account for
nonlinear material behavior and strength reduction due to liquefaction or strain softening.
In the pseudostatic method inertial forces induced by earthquake shaking are represented in terms of
pseudostatic accelerations, ah and av , and associated inertial forces, Fh and Fv:

ah av
Fh = W = kh W Fv = W = kvW (7.40)
g g

© 2003 by CRC Press LLC


0068_C07_fm Page 44 Wednesday, July 31, 2002 8:20 AM

7-44 Earthquake Engineering Handbook

FIGURE 7.48 Residual shear strength as a function of SPT penetration number. (From Seed, R.B. and L.F. Harder.
1990. “SPT-Based Analysis of Cyclic Pore Pressure Generation and Undrained Residual Strength,” H. Bolton Seed
Memorial Symposium. Used by permission of BiTech Publishers.)

TABLE 7.8 Correction of SPT Resistance for Estimating Residual Undrained


Shear Strength for Various Fine Contents According to the Procedures of Seed–Harder
and Stark–Mesri
Fines Content Ncor r Ncor r
(%) [Seed and Harder, 1990] [Stark and Mesri, 1992]

0 0 0
10 1 2.5
15 — 4
20 — 5
25 2 6
30 — 6.5
35 — 7
50 4 7
75 5 7

Data from Kramer [1996].

where ah and av are the horizontal and vertical accelerations, respectively, associated with a particular
level of earthquake shaking, and kh and kv are nondimensional pseudostatic earthquake coefficients. W
is the weight of the failed mass. These forces can be incorporated into any limit equilibrium analysis
procedure to determine an equivalent overall factor of safety. For example, in terms of the Ordinary
Method of Slices (Figure 7.49), the factor of safety is expressed as:

resisting force
∑ cb + [(W − F )cos α − F sin α]tan φ
v h

FS = = n
(7.41)
driving force (W − Fv ) sin α + Fh cos α
© 2003 by CRC Press LLC
0068_C07_fm Page 45 Wednesday, July 31, 2002 8:20 AM

Geotechnical and Foundation Aspects 7-45

Slice n
M number of slices
Fv
β Fh e
c, φ rfac
u
r es
W l u
fai
b α

FIGURE 7.49 Ordinary Method of Slices for slope stability analysis with pseudostatic seismic forces.

TABLE 7.9 Typical Values of Kh and FS for Use in Stability Calculations


kh FS Comments Source

0.10 >1.0 Major earthquake U.S. Army Corps of Engineers [1982]


0.15 >1.0 Great earthquake
0.05–0.15 Standard of practice; somewhat State of California
larger for critical conditions
0.15–0.25 >1.0 Standard of practice Japan
0.15 >1.15 With a 20% strength reduction Seed [1979a]
1/3 to 1/2 PGA >1.0 Marcuson and Franklin [1983]
1/2 PGA >1.0 With a 20% strength reduction Hynes-Griffin and Franklin [1984]

Source: Adapted from Abramson, L.W., T.S. Lee et al. 2002. Slope Stability and Stabilization Methods.
New York, John Wiley & Sons.

where c and φ are the Mohr-Coulomb strength parameters along the failure surface and the summation
is carried out over all M slices. Compared to the nonseismic case, F h clearly results in a reduction of the
FS. On the other hand, the effect of F v is less pronounced because it appears with the same sign in both
the numerator and the denominator. As a result, it is common to neglect F v altogether.
Most modern commercial limit equilibrium slope stability programs allow for this type of pseudostatic
analysis. The difficulty arises in selecting appropriate values of kh and FS. Because kh represents the inertial
shaking effects, it is reasonable to assume that it should be related in some fashion to the peak horizontal
acceleration amax (PHA). In general, slope deposits are compliant to various degrees and amax only occurs
over a very short period of time. Therefore, in practice kh is taken as a fraction of the maximum
acceleration. Considerable judgment is required in selecting appropriate values of kh. A number of
suggestions can be found in the literature [Seed, 1979a; U.S. Army Corps of Engineers, 1982; Marcuson
and Franklin, 1983; Hynes-Griffin and Franklin, 1984; Abramson et al., 2002] and some of these are
listed in Table 7.9. The value of kh is often prescribed in local codes. Although easy to conduct, the
pseudostatic approach is quite simplistic. It attempts to represent complex dynamic behavior in terms
of static forces. Stability is expressed in terms of an overall factor of safety. The implicit assumption is
that the soil is rigid-perfectly plastic and unchanging. This does not represent an appropriate approach
in cases where significant excess pore pressures may accumulate or where strength degradation due to
seismic loading is in excess of approximately 15% [Kramer, 1996].
Displacements associated with time-varying inertial forces can be estimated, to a first degree, with the
procedure proposed by Newmark [1965], which represents an extension of the pseudostatic approach.
An analogy is made between failure along a given sliding surface in a slope and a block initially resting
on an inclined surface (Figure 7.50). The block is subjected to horizontal inertial forces kh(t)W that

© 2003 by CRC Press LLC


0068_C07_fm Page 46 Tuesday, August 13, 2002 1:58 PM

7-46 Earthquake Engineering Handbook

A kh(t)W
kh(t)W

W
B W

C ay

FIGURE 7.50 Newmark sliding block method.

correspond to seismic motions propagating through the slope deposit. Displacement is initiated when
the sum of the downslope static and inertial forces equals the strength developed at the interface between
the block and the inclined plane. This condition occurs when the factor of safety is 1.0 and corresponds
to a yield coefficient ky and a yield acceleration ay = kyg. When the block is subjected to an interval with
acceleration larger than ay , it will begin to move relative to the plane (Figure 7.50). The corresponding
velocities and displacements can be obtained through integration of the acceleration record in excess of
ay . The assumption is that the sliding mass constitutes a rigid body. This assumption is only appropriate
for slopes where soils are very stiff or where the motion is of low frequency. Where this is not the case
careful consideration must be given in selecting an appropriate accelerogram. This may be done by
carrying out a site response analysis and determining acceleration series at various points along the
potential failure surface. An average horizontal equivalent acceleration (HEA) can be developed for use
with the Newark sliding block method. With reference to Figure 7.50, this would be computed as:

m Aa A (t ) + mB aB (t ) + mC aC (t )
HEA (t ) = (7.42)
m A + mB + mC

where m represents the mass of soil in each slice above the point where the acceleration response a(t) is
given. Conversely, a dynamic finite element analysis can be conducted to calculate average accelerations
over finite lengths of the potential failure surface based on integration of the time-dependent stresses.
These acceleration time series can in turn be used as input for the sliding block analysis. A number of
computer programs are available that can carry out such an analysis, including QUAD-4 [Idriss et al.,
1973] and FLUSH [Lysmer et al., 1975].
Makdisi and Seed [1978] developed a simplified procedure to estimate permanent horizontal displace-
ments of earth dams and embankments. These are determined with the aid of the charts in Figure 7.51.
The analysis is based on the dynamic response of embankments subjected to a range of ground motions
representing earthquakes of various magnitudes. Makdisi and Seed calculated the distribution of average
maximum acceleration with depth below the crest of the embankment (Figure 7.51A and B). Displace-
ments were estimated by comparing the acceleration at depth to the corresponding yield acceleration by
means of a Newmark-type sliding block analysis. The yield acceleration was taken as 80% of the undrained
shear strength of the soil. Results of the sliding block analysis are summarized in Figure 7.51C, from
which the horizontal displacement can be calculated for any failure surface extending a distance y below
the crest. To is the fundamental period of the dam, which can be obtained by means of an approximate
shear beam analysis [Gazetas, 1982] or other two-dimensional dynamic response modeling approach.
Although widely used, it is important to note that the Makdisi-Seed procedure is based on a limited set

© 2003 by CRC Press LLC


0068_C07_fm Page 47 Tuesday, August 13, 2002 1:58 PM

Geotechnical and Foundation Aspects 7-47

of case studies and, strictly speaking, should only be applied to dams and embankment slopes with
seismic motions corresponding to earthquake magnitudes in the range of 6.5 to 8.5 where the PGA at
the base of the embankment is at least 0.2 g.
A number of equivalent-linear procedures have been suggested to estimate approximate permanent
displacements in dams and slopes [Lee, 1974; Serff et al., 1976]. Their most appealing quality is that they
retain the simplicity of linear material behavior. They work well, provided that pore pressures remain
relatively low and that seismic motions do not induce excessive levels of material nonlinearity [Finn,
2001]. However, it is likely that in the future more reliance will be placed on stress-based finite element
and other similar numerical procedures that directly incorporate inelastic soil behavior. A sufficient
number of two- and three-dimensional analyses have been conducted using constitutive models ranging
from simple hysteretic to complex elasto-plastic models to validate this approach [Prevost et al., 1985;
Finn, 1988; Griffiths and Prevost, 1988; Marcuson et al., 1992]. Computer codes such as TARA-3 [Finn
et al., 1986], PLAXIS [Brinkgreve and Vermeer, 1988], and QUAKE/W [Quake/W, 2001] have been
developed to carry out these types of analysis.

7.7 Earthquake-Resistant Design of Retaining Walls


Soil pressures that act on retaining structures during earthquake shaking include both static and dynamic
components. Dynamic forces vary as the shaking proceeds and reflect not only the type of wall and soil
retained but also complex structure-interaction effects that in general are difficult to analyze. For example,
motion components that are close to the natural period of the soil-structure system can induce very large
transient pressures. Also, phase differences along the length of the retaining structure may induce signif-
icant shear forces and bending moments. In practice, however, most walls are designed using simplified
pseudostatic methods similar to those described previously for natural and constructed slopes. The
approach is to determine all static forces, along with pseudostatic seismic forces, and proceed with
conventional stability checks for overturning, sliding, bearing capacity, and overall stability.
Seismic effects are usually considered using the Mononobe–Okabe method [Mononobe and Matsuo,
1929; Okabe, 1929], which represents an extension of the Coulomb theory. The forces acting on active
and passive wedges of cohesionless soil are shown in Figure 7.52. The pseudostatic forces are given in
terms of the wedge weight W and the seismic coefficients kh and kv described earlier. For the active case
(Figure 7.52A), the total lateral force on the wall corresponds to the maximum value of Pae exerted by
any wedge with critical failure surface of inclination η.
This force can be expressed as:

γH 2 (1 − kv )K ae
1
Pae = (7.43)
2

where

K ae =
(
cos 2 φ − θ − β ) (7.44)
2
 
( )
 
1/ 2
 sin (δ + φ ) sin φ − α − β
( )
cos 2 θ cos β cos δ + θ + β 1 + 
( )
 
  cos δ + θ + β cos (α − θ)  
 
and

 k 
β = tan −1  h  (7.45)
 1 − kv 

© 2003 by CRC Press LLC


0068_C07_fm Page 48 Wednesday, July 31, 2002 8:20 AM

7-48 Earthquake Engineering Handbook

Y
H

(A)

F.E. Method
0.2

“Shear Slice”
(range for all data)
0.4

y/h

0.6

Average of
0.8 all data

10
0 0.2 0.4 0.6 0.8 1.0
kmax/ümax
(B)

10

1 M~8
M~8
M~7
7
U/kmax g To − seconds

6
0.1

M~6
0.01

0.001

(a) (b)
0.0001
0 0.2 0.4 0.6 0.8 1.0 0 0.2 0.4 0.6 0.8 1.0
ky/kmax ky/kmax
(C)
FIGURE 7.51 Simplified procedure to estimate permanent horizontal displacements of earth dams and embank-
ments. (From Makdisi, F.I. and H.B. Seed. 1978. “Simplified Procedure for Estimating Dam and Embankment
Earthquake-Induced Deformations,” J. Geotech. Eng. ASCE, 104(GT7), 849–867. Used by permission of the American
Society of Civil Engineers.)

© 2003 by CRC Press LLC


0068_C07_fm Page 49 Tuesday, August 13, 2002 1:58 PM

Geotechnical and Foundation Aspects 7-49

C
α
C
α

A A
Fv Fv
γ, φ
c=0
Fh R
θ Fh θ
φ
W φ W
Ppe
δ
R γ, φ
δ
c=0
Pae
η η
B B

(A) Active (B) Passive

FIGURE 7.52 Mononobe–Okabe method. Forces acting on active (A) and passive (B) wedges of cohesionless soil.

Note that if φ – α – β < 0, Equation 7.44 cannot be evaluated. This means that equilibrium is not
satisfied. For stability under seismic conditions it is therefore necessary that α ≤ φ – β . The location of
the combined wall force Pae can be obtained by separating Pae into static and dynamic components:

Pae = Pa + ∆ Pae (7.46)

The static force Pa acts at a height of H/3 above the base of the wall. According to Seed and Whitman
[1970], the dynamic component Pae acts at an approximate elevation of 0.6H above the base. Hence, the
location of Pae can be calculated as:

Pa ( H 3) + ∆Pae (0.6 H )
z= (7.47)
Pae

The critical failure surface for seismic conditions is flatter than for the static case. An expression for the
angle δ can be found in Kramer [1996]. The Mononobe–Okabe method for estimating Pae can also be
applied to c-ϕ backfill soil [Prakash and Saran, 1966].
For the passive case (Figure 7.52B), the total lateral force on the wall is given by:

γ H 2 (1 − kv )K pe
1
Ppe = (7.48)
2
where

K pe =
(
cos 2 φ + θ − β ) (7.49)
2
 
( 
)
1/ 2
  sin (δ + φ) sin φ + α − β  
2
(
cos θ cos β cos δ − θ + β 1 − ) ( ) 
  cos δ − θ + β cos (α − θ)  
 
The Mononobe–Okabe method is subject to the limitations of the Coulomb theory and to the same
uncertainties associated in selecting appropriate coefficients kh and kv as was discussed earlier. This
method should not be used for soils that may liquefy or otherwise may lose strength due to the shaking.
Wood [1973] showed that where the principal energy of the input motions approaches the fundamental
frequency of the unrestrained backfill ( fo = Vs /4H), dynamic amplification becomes an important factor,
which is not considered in any of the analysis procedures described in this section.

© 2003 by CRC Press LLC


0068_C07_fm Page 50 Wednesday, July 31, 2002 8:20 AM

7-50 Earthquake Engineering Handbook

Elastic soil

H
L

Unyielding wall

1.2
ν = 0.5
ν = 0.4
1.0

0.8 ν = 0.3 0.8


ν = 0.5
ν = 0.2 ν = 0.4
Fp 0.6 0.6

0.4 Fm 0.4
ν = 0.3
ν = 0.2
0.2 0.2

0 0
0 2 4 6 8 10 0 2 4 6 8 10
L/H L/H

Dimensionless thrust factor for Dimensionless moment factor


various geometries and soil Poisson’s ratio for various geometries and soil Poisson’s ratio
values. values.

FIGURE 7.53 Dynamic loads on rigid retaining walls. (Modified from Wood, J. 1973. “Earthquake-Induced Soil
Pressures on Structures,” Report EERL 73-05, California Institute of Technology, Pasadena, p. 311.)

Very large gravity walls and other retaining structures that are restrained may not yield sufficiently to
reach active or passive plastic equilibrium states. Wood [1973] presents a method to calculate dynamic
lateral forces and overturning moments for a rigid wall by assuming that the backfill soil behaves in a
homogeneous, linearly elastic fashion, and that the seismic excitation is due to a uniform, constant,
horizontal acceleration throughout the backfill. Wood actually considered two rigid walls separated by
a distance L of sufficient magnitude to avoid interaction between the two of them (Figure 7.53).
The dynamic force and moment are calculated as:

ah
∆Poe = γ H 2 F (7.50)
g p

ah
∆M oe = γ H 3 F (7.51)
g m

© 2003 by CRC Press LLC


0068_C07_fm Page 51 Wednesday, July 31, 2002 8:20 AM

Geotechnical and Foundation Aspects 7-51

where ah is the amplitude of the motion. The dimensionless factors F p and F m are shown in Figure 7.53.
The point of application of ∆Poe, above the base of the wall, is given by:

∆M oe
z= ≈ 0.63H (7.52)
∆Poe

The previous methods assume that the backfill is not immersed in water. Although walls are typically
built with provisions for proper drainage, this is not possible for retaining structures where water is
found on either side, as is the case for many harbor and other shore-front structures. Extensive damage
to such structures was observed during recent earthquakes in Turkey, Taiwan, and Japan and in many
cases were attributed to dynamic water effects. Water influences the inertial forces in the backfill and
may result in substantial hydrodynamic pressures. Also, a loose, saturated granular backfill may liquefy
and cause undesirable settlements. If the permeability of the soil is sufficiently small (typically less than
about 10–3 cm/sec), the solid and fluid fractions are likely to move in unison, whereas if the permeability
is very large, the fluid will tend to move independently of the solid structure. In the former case, the
inertial forces are dependent on the total unit weight of the soil, whereas in the latter case they will be
a function of the submerged unit weight. If the type of backfill is such that water is restrained by the
soil skeleton, the Monotobe–Okabe method can be modified to give the active total soil force on the wall
by replacing the soil unit weight in Equation 7.43 according to:

γ = γ b (1 − ru ) (7.53)

where r u is the pore pressure ratio, defined as the pore pressure divided by the effective confining pressure,
and γb is the submerged unit weight. The seismic coefficient given in Equation 7.45 now becomes:

 γ sat kh 
β = tan −1   (7.54)
 γ b (1 − ru ) (1 − kv ) 

where γsat is the saturated unit weight of the soil. In addition to the hydrostatic water pressure in the
absence of shaking, an additional equivalent hydrostatic term must be considered which is calculated
using an equivalent fluid unit weight:

γ w −eq = γ w + ru γ b (7.55)

If the backfill is partially submerged to a height λH above the base of the wall (λ < 1.0), it is necessary
to use an average soil unit weight:

(
γ = λ2 γ sat + 1 − λ2 γ d ) (7.56)

where γd is the dry unit weight of the soil. The additional equivalent hydrostatic term calculated with the
equivalent fluid unit weight in Equation 7.55 also must be included for partially submerged backfill.
Permanent displacements of yielding walls are often a greater concern than outright failure because
excessive deformations can severely limit their intended function. Richards and Elms [1979] use
Newmark’s sliding block method to estimate permanent lateral displacement of gravity walls. Consider
a gravity wall acted upon by a Monotobe–Okabe lateral force Pae during seismic shaking, as shown in
Figure 7.54. Lateral displacement will be initiated on the verge of active failure, at which point plastic
equilibrium dictates that:

© 2003 by CRC Press LLC


0068_C07_fm Page 52 Wednesday, July 31, 2002 8:20 AM

7-52 Earthquake Engineering Handbook

θ
Pae
δ
khW

FIGURE 7.54 Forces acting on gravity wall


during seismic shaking. φb

N = W + Pae sin (δ + θ) (7.57)

T = khW + Pae cos (δ + θ) (7.58)

T = N tan φb (7.59)

where φb is the friction angle between the base and the foundation soil. These expressions can be combined
to obtain:

ay Pae cos (δ + θ) − Pae sin (δ + θ) tan φb


kh = = tan φb − (7.60)
g W

The term ay represents the yield acceleration, beyond which lateral displacement occurs. Using the
results of Newmark [1965] and Franklin and Chang [1977], Richards and Elms developed the following
relationship to estimate permanent lateral displacement:

v p2 a3p
d perm = 0.087 (7.61)
a3y

In this equation vp is the peak velocity and ap the peak acceleration at the wall location.
This type of analysis should not be used indiscriminately because it is rather simplistic and ignores
several effects that may play an important role in some cases. Only accelerations in the direction normal
to the length of the wall are considered. Neither vertical nor horizontal acceleration components in the
direction of the length of the wall are accounted for. Also, the backfill and the wall are assumed to act
as one and no amplification of input motions occurs in the retained soil. Only lateral sliding displace-
ments are estimated but not tilting or vertical settlements. Whitman and Liao [1985] considered some
of these effects and estimated errors associated with them.
Kramer [1996] presents procedures for the design of various types of retaining structures that are
based on the pseudostatic methods described in this section. A more rigorous approach to the design
and analysis of retaining structures is possible by means of the finite element method. In fact, the finite
element method is quickly becoming the method of choice for the design of all but the simplest of
retaining structures.

© 2003 by CRC Press LLC


0068_C07_fm Page 53 Wednesday, July 31, 2002 8:20 AM

Geotechnical and Foundation Aspects 7-53

7.8 Soil Remediation Techniques for Mitigation


of Seismic Hazards
If seismic hazards are deemed to be unacceptably high because of poor soil conditions, it is often possible
to achieve improved seismic performance through the use of one or more soil remediation techniques.
Poor performance is the result of (1) inadequate strength, (2) low stiffness, or (3) insufficient drainage.
Many remediation techniques have evolved over the years, mostly through trial and error, aimed at
improving at least one of these properties. When selecting one or more mitigation methods, it is important
to consider the effectiveness of the remediation approach for the particular situation at hand, cost,
environmental consequences, regulatory requirements, and technical feasibility. Also, careful assessment
of the degree of ground improvement achieved is essential. The subject of soil remediation is quite
extensive and a number of excellent sources and case studies are available in the literature [Hausmann,
1990; Hryciw, 1995; Mitchell et al., 1995; Schaefer, 1997]. The purpose of this section is to highlight the
most promising techniques for improving seismic performance.
Excavation and replacement may be a cost-effective solution for sufficiently shallow deposits. Placing
a structure at depth may bypass undesirable surface soils, although costs and construction difficulties
increase rapidly with depth and with excavation below the water table, particularly in high-permeability
soils. Surrounding deposits that have not been modified may still cause problems as lifelines and other
connecting structures may be damaged during an earthquake.
Compaction can be accomplished using a variety of techniques that are aimed at increasing the density
of soil, thereby resulting in improved stiffness, strength, and liquefaction resistance. Vibrocompaction is
most effective for clean, loose cohesionless soils with less than about 15% silt and less than about 3%
clay content. It is achieved by vibration of the head of the vibration probe as it is withdrawn. Because
compaction occurs only within a short range of the probe, the procedure must be repeated at regular
spacings on the order of 5 to 10 ft. Of course, the spacing depends on the size of the probe and the soil
type. During the past few years, larger and more powerful vibrators have been introduced, which allow
larger spacings and deeper penetration (in some cases, up to 120 ft). When compaction is achieved by
horizontal motion of the vibrator, it is referred to as vibroflotation. Vibratory techniques also exist that
induce vertical vibration, such as Terra-Probe, Vibro-Wing, and Tri-Star or Y-Probe methods [Hryciw,
1995]. Wightman [1991] presents an overview of this technique. Dynamic compaction involves dropping
a heavy weight from a large distance. The high energy upon impact is provided by heavy steel or concrete
units (6 to 35 tons) that freefall from distances up to 100 ft or more. It often represents an alternative
to vibrocompaction, especially for uncontrolled fills, municipal solid waste deposits, coal mine spoils,
and other loose soils. Soils with significant amounts of fines (20% or more) can in some cases be densified
quite effectively. The depth of improvement is related to the tamper weight and drop height but may
reach up to 30 ft or more. A good reference on dynamic compaction is the FHWA publication by Lukas
[1986], which was updated in 1996 as FHWA Geotechnical Engineering Circular No. 1. Blast densification
is another high-energy ground improvement technique that achieves densification by destroying existing
soil structure and forcing soil grains into a tighter configuration as a result of shock waves produced by
the blast. Charges are placed in predrilled or jetted holes. The size of the charge must be selected carefully
so that it is sufficiently large to be effective but not too intense to cause excessive vibrations that may
cause damage to nearby structures. Liquefaction may develop and may have to be controlled by proper
drainage means. Because of the potential of undesirable effects in surrounding areas, blast densification
has not seen the same degree of use as the previous techniques. However, it has been shown effective in
densifying soils to depths of approximately 130 ft [Narin van Court and Mitchell, 1995]. Compaction
piles achieve densification by displacing the soil as the piles are driven into the ground. Because they
typically densify the soil to distances on the order of a only few pile diameters, they must be placed close
together to be effective. Improvements have been noted to depths of about 60 ft [Marcuson et al., 1991].
Grouting involves the injection of various grouting agents into the soil. Compaction grouting has been
shown to be effective for mitigation of liquefaction potential [Graf, 1992; Boulanger and Hayden, 1995].
The technique consists of injecting a soil-cement grout of sufficient plasticity and friction under pressure,

© 2003 by CRC Press LLC


0068_C07_fm Page 54 Wednesday, July 31, 2002 8:20 AM

7-54 Earthquake Engineering Handbook

which displaces and densifies soil in a controlled fashion. Although the technique is widely used for a
number of purposes, there is little in terms of a rational design methodology. Instead, the method has
progressed based almost entirely on trial and error and a few empirical observations. Research is now
under way to establish optimum grout characteristics, injection pressures, and effective pumping rates
vs. soil characteristics [Schaefer, 1997]. Advances have been made recently in terms of equipment and
monitoring, and particularly in terms of evaluating the degree of improvement through seismic testing.
Whereas compaction grouting results in discrete bulbs of grout in the soil, permeation grouting uses
low viscosity grouts that are able to penetrate into individual voids with minimal disturbance to the soil
structure. The types of grouts used range from high-slump cements to various gels of very low viscosity,
depending primarily on the void characteristics of the soil [Graf, 1992]. In jet grouting, a high-pressure
fluid is used to erode soil in a predrilled hole and replace it with an engineered soil-grout mix to form
a solid element sometimes referred to as soilcrete or grout column. The dimensions of the grouted cavities
are controlled by the injection pressure, the type and operation of the injection nozzle, and the erosion
susceptibility of the soil. Jet grouting is most successful in cohesionless soils and can be performed as
deep as predrilled holes can be provided. This technique has been used successfully as a liquefaction
countermeasure [Hayden, 1994]. Another soil improvement technique that can be used to mix under-
performing soil with grouts and admixtures is to use large rotary augers to churn up soil and blend in
cementitous agents that result in increased stiffness and strength. Soil-mixing can be used to provide
support for overlaying structures or to reduce liquefaction hazards [Schaefer, 1997].
Vibrostone columns have been used to improve soils prone to liquefaction since the 1970s [Dobson,
1987]. Construction is accomplished by introducing a vibratory probe into the ground, which displaces
the soil laterally through vibratory motion and therefore induces densification in the surrounding
volume. The void that is created is backfilled with stone. The resulting column and surrounding soil
provide for higher stiffness and strength. Also, the damaging effects of liquefaction are reduced because
the stone columns provide a relief path for excess pore pressures to dissipate. A review of the performance
of vibrostone columns for reduction of soil liquefaction is presented by Baez [1995]. Wick drains can
also be used to dissipate excess pore pressures generated during earthquake shaking. They consist of
either properly graded sand and gravel drains or of prefabricated geosynthetic materials.
Figure 7.55 illustrates the general soil particle size ranges for applicability of various stabilization
techniques. An excellent review of liquefaction remediation techniques is presented in PHRI [1997].

Defining Terms
Amplification — Increase in ground motion due to the presence of soil deposits, usually expressed in
terms of the ratio of ground surface to bedrock motion.
Attenuation — Rate of seismic ground motion decrease with distance.
Backbone curve — Nonlinear stress-strain relationship for a soil that is loaded monotonically.
Bracketed duration — Time between the first and last exceedance of a given threshold acceleration
during a seismic event.
Cyclic resistance ratio — Cyclic stress ratio above which liquefaction is triggered.
Cyclic strength — Shear stress during cyclic loading at which the resulting deformations are considered
excessive.
Cyclic stress ratio — Ratio of earthquake-induced equivalent shear stress amplitude to effective over-
burden stress.
Damping curve — Relationship between viscous damping coefficient and shear strain for nonlinear
soils.
Epicenter — Projection on the surface of the Earth directly above the location where the initial seismic
disturbance occurred.
Equivalent damping ratio — Ratio between the dissipated and stored energies within a hysteretic loop
at a given shear strain.

© 2003 by CRC Press LLC


0068_C07_fm Page 55 Wednesday, July 31, 2002 8:20 AM

Geotechnical and Foundation Aspects 7-55

Gravel Sand Silt Clay

Vibro-compaction

Blasting

Particulate Grout

Chemical Grout

Displacement Grout

Pre-compression

Dynamic Deep Compaction

Electro-osmosis

Reinforcement (tension, compression, shear)

Thermal Treatment

Admixtures

10 1.0 0.1 0.01 0.00 0.0001


1
Particle Size (mm)

FIGURE 7.55 Applicable grain size ranges for different stabilization methods. (Modified from Mitchell, J.K. 1981.
“Soil Improvement: State-of-the-Art,” in Proc. Tenth International Conference on Soil Mechanics and Foundation
Engineering, Stockholm, American Society of Civil Engineers, New York.)

Equivalent-linear soil model — A stress-strain model that uses elastic shear modulus and damping
ratio parameters that are functions of shear strain. These parameters are updated iteratively
during each incremental step until they match the target values corresponding to the level of
strain in the soil.
Flow failures — Soil failure due to liquefaction involving large lateral displacements that occur when
the static shear stress along a potential failure plane is larger than the shear strength of the
liquefied soil.
Intensity — Strength of shaking from a particular earthquake at a given location.
Lateral spreading — Small to moderate lateral displacements associated with liquefaction that occur
due to seismic shaking when the static shear stress is less than the shear strength of the liquefied
soil.
Liquefaction — A process in which the soil loses shear strength and approaches the state of a liquid due
to a transient accumulation of excess pore pressures.
Magnitude — A measure of the energy released at the source of the earthquake.
Magnitude scaling factors — Factors to be applied to the cyclic resistance ratio obtained from the
standard penetration test, cone penetration test, and shear wave velocity liquefaction charts for
earthquakes of magnitude other than 7.5.
Modulus reduction curve — Relationship expressing the rate at which the maximum shear modulus
of soil decreases with shear strain due to nonlinear effects.
Near-field — Within a distance equal to the dimension of the fault section involved in the earthquake.
Nonlinear cyclic model: — A stress-strain model that is truly nonelastic and that models the nonlin-
earity through an explicit relationship involving nonelastic material parameters.
Peak ground acceleration (PGA) — Maximum recorded acceleration amplitude.

© 2003 by CRC Press LLC


0068_C07_fm Page 56 Tuesday, August 13, 2002 1:58 PM

7-56 Earthquake Engineering Handbook

Pseudostatic approach — Stability analysis procedure in which inertial forces caused by earthquake
shaking are approximated by equivalent static forces that are a function of peak ground motion
acceleration and soil weight.
Residual shear strength — Soil shear strength after seismic shaking has stopped.
Response spectrum — A plot of maximum acceleration, velocity, or displacement for a single-degree-
of-freedom oscillator as a function of system period, for a given input motion and system
damping (typically 5%).
Sand boils — Sand and silt mounds deposited by spouting from crater-like vents as excess pore pressure
dissipates from buried deposits that have liquefied.
Seismic hazards — Possible types of damage from earthquake shaking.
Seismic waves — Waves originating from an earthquake event and traveling through geologic media,
including compressional p-waves, shear S-waves, surface Love waves, and long-period Rayleigh
waves.
Site-specific ground response analysis — Ground motion analysis that involves site-specific material
distributions, dynamic soil response, and boundary conditions. It is usually carried out using
a numerical approach.
Strong motion — Seismic shaking that is of sufficient intensity to have a significant effect on engineering
structures.
Tsunami — Large tidal wave that follows from the sudden displacement of the seafloor or a submarine
landslide, usually caused by an offshore earthquake.
Yield acceleration — Acceleration corresponding to a pseudostatic safety factor equal to 1, above which
permanent deformations accumulate.

References
Abrahamson, N.A. and K.M. Shedlock. 1997. “Overview,” Seismol. Res. Lett., 68(1), 9–23.
Abrahamson, N.A. and P.G. Somerville. 1996. “Effects of the Hanging Wall and Footwall on Ground
Motions Recorded during the Northridge Earthquake,” Bull. Seismol. Soc. Am., 86, S93–S99.
Abramson, L.W., T.S. Lee, et al. 2002. Slope Stability and Stabilization Methods, John Wiley & Sons, New
York.
Aki, K. 1988. “Local Site Effects on Strong Ground Motion,” in Earthquake Engineering and Soil Dynamics,
Vol, II, Recent Advances in Ground Motion Evaluation, Geotechnical Special Publication No. 20,
American Society of Civil Engineers, New York.
Andrews, D.C.A. and G.R. Martin. 2000. “Criteria for Liquefaction of Silty Soils,” in Proc. 12th World
Conference on Earthquake Engineering, Auckland, New Zealand.
Andrus, R.D. and K.H. Stokoe. 1997. “Liquefaction Resistance Based on Shear Wave Velocity,” in Proc.
NCEER Workshop on Evaluation of Liquefaction Resistance of Soils, Salt Lake City, NV, Multidis-
ciplinary Center for Earthquake Engineering Research, Buffalo, NY.
Arnold, C. and R. Reitherman. 1982. Building Configuration and Seismic Design. John Wiley & Sons, New
York.
Baez, J.I. 1995. “A Design Model for the Reduction of Soil Liquefaction by Vibro-Stone Columns,”
University of Southern California, Los Angeles, p. 207.
Bardet, J.-P., N. Mace, et al. 1999. “Liquefaction-Induced Ground Deformation and Failure,” University
of California, Department of Civil Engineering, p. 125.
Bolt, B.A. 1969. “Duration of Strong Motion,” in Proc. 4th World Conference on Earthquake Engineering,
Santiago, Chile.
Boore, D.M., W. Joyner, et al. 1997. “Empirical Near-Source Attenuation Relationships for Horizontal
and Vertical Components of Peak Ground Acceleration, Peak Ground Velocity, and Pseudo-Abso-
lute Acceleration Response Spectra,” Seismol. Res. Lett., 68(1), 154–179.
Boulanger, R.W. and R.F. Hayden. 1995. “Aspects of Compaction Grouting of Liquefiable Soils,” ASCE J.
Geotech. Eng., 12(121), 844–855.

© 2003 by CRC Press LLC


0068_C07_fm Page 57 Wednesday, July 31, 2002 8:20 AM

Geotechnical and Foundation Aspects 7-57

Brinkgreve, R.B.J. and P.A. Vermeer. 1988. “PLAXIS, Finite Element Code for Soil and Rock Analysis,
Version 7, Balkema, Rotterdam.
BSSC (Building Seismic Safety Council). 1997. NEHRP Recomme nded Provisions for Seismic Regulat ions
for New Building s and Othe r Structures, National Institute of Building Sciences, p. 335.
Castro, G. and S.J. Poulos. 1977. “Factors Affecting Liquefaction and Cyclic Mobility,” J. Geotech. Eng.
Div. ASCE, 106(GT6), 501–506.
Cetin, K.O. and R.B. Seed. 2000. “Nonlinear Shear Mass Participation Factor (Rd) for Cyclic Shear Stress
Ratio Evaluation,” University of California, Berkeley.
Chopra, A.K. 2001. Dynamic s of Structures, Theory and Applicat ions to Earthquak e Engineering. Prentice-
Hall, Upper Saddle River, NJ.
Cruden, D.M. and D.J. Varnes. 1996. “Landslide Types and Processes,” in Landslides, Investigation and
Mitigation, Transportation Research Board Special Report 247, pp. 36–75, A.K. Turner and R.L.
Schuster, Eds., National Academy Press, Washington, D.C.
Dobson, T. 1987. Case Histories of the Vibro Systems to Minimiz e the Risk of Liquefaction, Geotechnical
Special Publication No. 12, American Society of Civil Engineers, New York, pp. 167–183.
Faccioli, E. 1991. “Seismic Amplification in the Presence of Geologic and Topographic Irregularities,” in
Proc. Second International Conference on Recent Advances in Geotechnical Earthquake Engineer-
ing and Soil Dynamics, St. Louis, MO.
Finn, W.D.L. 1988. “Dynamic Analysis in Geotechnical Engineering,” in Earthquak e Engineering and Soil
Dynamic s, Vol. II: Recent Advances in Ground-Motion Evaluat ion, Geotechnical Special Publication
20, American Society of Civil Engineers, New York.
Finn, W.D.L. 1991. “Geotechnical Engineering Aspects of Microzonation,” in Proc. Fourth International
Conference on Microzonation, Earthquake Engineering Research Institute, Stanford University,
Palo Alto.
Finn, W.D.L. 2000. “State-of-the-Art of Geotechnical Earthquake Engineering Practice,” Soil Dyn. Earth-
quake Eng., 20, 1–15.
Finn, W.D.L. 2001. “Earthquake Engineering,” in Geotechnical and G eoenvironmental E ngineering Hand-
book, R.K. Rowe, Ed., Kluwer Academic, Boston, pp. 615–659.
Finn, W.D.L., P.M. Byrne, et al. 1996. “Some Geotechnical Aspects of the Hyogo-ken-Nanbu (Kobe)
Earthquake of January 17, 1995,” Can. J. Civil Eng., 23(3), 778–796.
Finn, W.D.L., R.H. Ledbetter, et al. 1994. Liquefaction in Silty Soils: Design and Analysis, Geotechnical
Publication No. 44, American Society of Civil Engineers, New York, pp. 51–76.
Finn, W.D.L., M. Yogendrakumar, et al. 1986. “TARA-3: A Problem for Non-Linear Static and Dynamic
Effective Stress Analysis,” Soil Dynamics Group, University of British Columbia, Vancouver.
Franklin, A.G. and F.K. Chang. 1977. “Permanent Displacements of Earth Embankments by Newmark
Sliding Block Analysis,” U.S. Army Corps of Engineers, Waterways Experiment Station, Vicksburg,
MS.
Gazetas, G. 1982. “Shear Vibrations of Vertically Inhomogeneous Earth Dams,” Int. J. Num. Analyt. Meth.
Geomech., 6(1), 219–241.
Geist, E.L. 2000. “Origin of the 17 July 1998 Papua New Guinea Tsunami: Earthquake or Landslide?”
Seismol. Res. Lett., 71(3), 344–351.
Graf, E.D. 1992. Earthquak e Support Grouting in Sands , Geotechnical Special Publication No. 30, American
Society of Civil Engineers, New York, pp. 265–274.
Graves, R.W. 1993. “Modeling Three-Dimensional Site Response Effects in the Marina District Basin,
San Francisco, California,” Bull. Seismol. Soc. Am., 83: 1042–1063.
Graves, R.W., A. Pitarka et al. 1998. “Ground Motion Amplification in the Santa Monica Area: Effects of
Shallow Basin Edge Structure,” Bull. Seismol. Soc. Am., 88(5), 1224–1242.
Griffiths, D.V. and J.H. Prevost. 1988. “Two- and Three-Dimensional Finite Element Analysis of the Long
Valley Dam,” National Center for Earthquake Engineering Research, University of New York,
Buffalo.

© 2003 by CRC Press LLC


0068_C07_fm Page 58 Tuesday, August 13, 2002 1:59 PM

7-58 Earthquake Engineering Handbook

Harder, L.F. and R. Boulanger. 1997. “Application of Ksigma and Kalpha Correction Factors,” in Proc. NCEER
Workshop on Evaluation of Liquefaction Resistance of Soils, Salt Lake City, NV, Multidisciplinary
Center for Earthquake Engineering Research, Buffalo, NY.
Harder, L.F. and H.B. Seed. 1986. “Determination of Penetration Resistance for Coarse-Grained Soils
Using the Becker Hammer Drill,” Earthquake Engineering Research Center, University of Califor-
nia, Berkeley.
Hausmann, M.R. 1990. Engineering Principles of Ground Modification. McGraw-Hill, New York.
Hayden, R.F. 1994. “Utilization of Liquefaction Countermeasures in North America,” in Proc. Fifth U.S.
National Conference on Earthquake Engineering, Chicago.
Hryciw, R.D., Ed. 1995. Soil Improvement for Earthquake Hazard Mitigation, Geotechnical Special Pub-
lication No. 49, American Society of Civil Engineers, New York.
Hynes-Griffin, M.E. and A.G. Franklin. 1984. “Rationalizing the Seismic Coefficient Method,” U.S. Army
Corps of Engineers, Waterways Experiment Station 21, Vicksburg, MS.
ICBO (International Conference of Building Officials). 1997. Uniform Building Code, International Con-
ference of Building Officials, Whittier, CA.
ICC (International Code Council). 2000. International Building Code, International Code Council, Falls
Church, VA.
Idriss, I.M. 1990. “Response of Soft Soil Sites during Earthquakes,” H. Bolton Seed Memorial Symposium,
BiTech Publishers.
Idriss, I.M., J. Lysmer, et al. 1973. “QUAD-4: A Computer Program for Evaluating the Seismic Response
of Soil Structures by Variable Damping Finite Element Procedures,” Earthquake Engineering
Research Center, University of California, Berkeley.
Kramer, S.L. 1996. Geotechnical Earthquake Engineering, Prentice-Hall, Upper Saddle River, NJ.
Lee, K.L. 1974. “Seismic Permanent Deformations in Earth Dams,” School of Engineering and Applied
Science, University of California, Los Angeles.
Lee, K.L. and J.A. Focht. 1976. “Strength of Clay Subjected to Cyclic Loading,” Mar. Geotechnol., 1(3),
165–188.
Liao, S.S.C. and R.V. Whitman. 1986. “Overburden Correction Factors for SPT in Sand,” J. Geotech. Eng.
ASCE, 112(3), 373–377.
Lukas, R.G. 1986. Dynamic Compaction for Highway Construction, Vol. 1. Design and Construction Guide-
lines, Federal Highway Administration.
Lysmer, J., T. Udaka, et al. 1975. “FLUSH: A Computer Program for Approximate 3-D Analysis of Soil-
Structure Interaction Problems,” Earthquake Engineering Research Center, University of California,
Berkeley.
Makdisi, F.I. and H.B. Seed. 1978. “Simplified Procedure for Estimating Dam and Embankment Earth-
quake-Induced Deformations,” J. Geotech. Eng. ASCE, 104(GT7), 849–867.
Marcuson, W.F. and A.G. Franklin. 1983. Seismic Design, Analysis and Remedial Measures to Improve the
Stability of Existing Earth Dams, Corps of Engineers Approach, Seismic Design of Embankments and
Caverns, American Society of Civil Engineers, New York.
Marcuson, W.F., P.F. Hadala, et al. 1991. Seismic Rehabilitation of Earth Dams, Geotechnical Publication
No. 35, American Society of Civil Engineers, New York, pp. 430–466.
Marcuson, W.F., M.E. Hynes, et al. 1992. “Seismic Stability and Permanent Deformation Analysis: The
Last Twenty Five Years,” in ASCE Specialty Conference on Stability and Performance of Slopes and
Embankments, Vol. II, Geotechnical Special Publication No. 31, American Society of Civil Engi-
neers, New York.
Mitchell, J.K. 1981. “Soil Improvement: State-of-the-Art,” in Proc. Tenth International Conference on
Soil Mechanics and Foundation Engineering, Stockholm, American Society of Civil Engineers,
New York.
Mitchell, J.K., C.D.P. Baxter, et al. 1995. “Performance of Improved Ground During Earthquakes,” in Soil
Improvement for Earthquake Mitigation, American Society of Civil Engineers, New York.

© 2003 by CRC Press LLC


0068_C07_fm Page 59 Tuesday, August 13, 2002 1:59 PM

Geotechnical and Foundation Aspects 7-59

Molas, G.L. and F. Yamazaki. 1995. “Attenuation of Earthquake Ground Motion in Japan Including Deep
Focus Events,” Bull. Seismol. Soc. Am., 85(5), 1343–1358.
Molas, G.L. and F. Yamazaki. 1996. “Attenuation of Response Spectra in Japan Using New JMA Records,”
Bull. Earthquake Resistant Structure Research Center, Institute of Industrial Science, University of
Tokyo (29), 115–28.
Mononobe, N. and H. Matsuo. 1929. “On the Determination of Earth Pressures During Earthquakes,”
in Proc. World Engineering Conference.
Narin van Court, W.A. and J.K. Mitchell. 1995. New Insights into Explosive Compaction of Loose, Saturated,
Cohesionless Soils, Special Geotechnical Publication No. 49, American Society of Civil Engineers,
New York, pp. 51–65.
NCEER (National Center for Earthquake Engineering Research). 1992. Case Studies of Liquefaction and
Lifeline Performance during Past Earthquakes. Vol. 1, Japanese Case Studies; Vol. 2, United States
Case Studies, National Center for Earthquake Engineering Research, State University of New York,
Buffalo.
NCEER (National Center for Earthquake Engineering Research). 1997. Proceedings of the NCEER Work-
shop on Evaluation of Liquefaction Resistance of Soils, Summary Report, T.L. Youd and I. Idriss, Eds.,
National Center for Earthquake Engineering Research, State University of New York, Buffalo.
Newmark, N.M. 1965. “Effects of Earthquakes on Dams and Embankments,” Geotechnique, 15(2),
139–160.
Newmark, N.M. and W.J. Hall. 1982. Earthquake Spectra and Design, Earthquake Engineering Research
Institute, University of California, Berkeley, p. 103.
Okabe, S. 1929. “General Theory of Earth Pressures,” J. Jpn. Soc. Civil Eng., 12(1).
O’Rourke, T.D. 1992. The Loma Prieta, California, Earthquake of October 17, 1989: Marina District, United
States Geological Service, Government Printing Office, Washington, D.C.
Otani, S. 2000. “New Seismic Design Provisions in Japan,” Pacific Earthquake Engineering Research
Center, University of California, Berkeley, pp. 3–14.
PHRI (Port and Harbor Research Institute). 1997. Handbook on Liquefaction Remediation of Reclaimed
Land, Port and Harbour Research Institute, Japan, Brookfield, VT.
Poulos, S.J., G. Castro, et al. 1985. “Liquefaction Evaluation Procedure,” J. Geotech. Eng. ASCE, 111(6),
772–792.
Prakash, S. and S. Saran. 1966. “Static and Dynamic Earth Pressure Behind Retaining Walls,” in Proc.
Third Symposium on Earthquake Engineering, Roorkee, India.
Prevost, J.H., A.M. Abdel-Ghaffar, et al. 1985. “Nonlinear Dynamic Analysis of Earth Dams: A Compar-
ative Study,” J. Geotech. Eng. ASCE, 111(2), 882–897.
Quake/W. 2001. “Quake/W for Finite Element Dynamic Earthquake Analysis: Users’ Guide,” GEO-SLOPE
International Ltd., Calgary.
Richards, R. and D.G. Elms. 1979. “Seismic Behavior of Gravity Retaining Walls,” J. Geotech. Eng. ASCE,
105(GT4), 449–464.
Robertson, P.K. and C.E. Fear. 1995. “Liquefaction of Sand and its Evaluation,” in Proc. 1st International
Conference on Earthquake Geotechnical Engineering, Tokyo.
Robertson, P.K. and C.E. Wride. 1997. “Cyclic Liquefaction and its Evaluation Based on the SPT and
CPT,” in Proc. NCEER Workshop on Evaluation of Liquefaction Resistance of Soils, Salt Lake City,
NV, Multidisciplinary Center for Earthquake Engineering Research, Buffalo, NY.
Rodriguez-Marek, A., J.D. Bray, et al. 1999. “Characterization of Site Response, General Site Categories,”
Pacific Earthquake Engineering Research Center, University of California, Berkeley, p. 134.
Romo, M.P. and H.B. Seed. 1986. “Analytical Modeling of Dynamic Soil Response in the Mexico Earth-
quake of September 19, 1985,” in Proc. ASCE International Conference on the Mexico Earthquakes
1985, Mexico City, American Society of Civil Engineers, New York.
Schaefer, V.R., Ed. 1997. Ground Improvement, Ground Reinforcement, Ground Treatment, Geotechnical
Special Publication No. 69, American Society of Civil Engineers, New York.

© 2003 by CRC Press LLC


0068_C07_fm Page 60 Wednesday, July 31, 2002 8:20 AM

7-60 Earthquake Engineering Handbook

Seed, H.B. 1979a. “Considerations in the Earthquake-Resistant Design of Earth and Rockfill Dams,”
Geotechnique, 29(3), 215–263.
Seed, H.B. 1979b. “Soil Liquefaction and Cyclic Mobility for Level Ground During Earthquakes,”
J. Geotech. Eng. ASCE, 105(GT2), 201–255.
Seed, H.B. 1983. “Earthquake Resistant Design of Earth Dams,” in Proc. Symposium on Seismic Design
of Earth Dams and Caverns, New York.
Seed, H.B. 1986. “Design Problems in Soil Liquefaction,” J. Geotech. Eng. ASCE, 113(8), 827–845.
Seed, R.B. and L.F. Harder. 1990. “SPT-Based Analysis of Cyclic Pore Pressure Generation and Undrained
Residual Strength,” H. Bolton Seed Memorial Symposium, BiTech Publishers.
Seed, H.B. and I.M. Idriss. 1970. “Soil Moduli and Damping Factors for Dynamic Response Analysis,”
Earthquake Engineering Research Center, University of California, Berkeley.
Seed, H.B. and I.M. Idriss. 1971. “Simplified Procedure for Evaluating Soil Liquefaction Potential,” J. Soil
Mech. Found. Div. ASCE, 97(SM9), 1249–1273.
Seed, H.B. and I.M. Idriss. 1982. Ground Motions and Soil L iquefaction during Earthquak es, Monogr. 5,
Earthquake Engineering Research Institute, University of California, Berkeley.
Seed, H.B. and R.V. Whitman. 1970. “Design of Earth Retaining Structures for Dynamic Loads,” in Proc.
ASCE Specialty Conference on Lateral Stresses in the Ground and Design of Earth Retaining
Structures, American Society of Civil Engineers, New York.
Seed, R.B., K.O. Cetin, et al. 2001. “Recent Advances in Soil Liquefaction Engineering and Seismic Site
Response Evaluation,” in Proc. Fourth International Conference on Recent Advances in Geotech-
nical Earthquake Engineering and Soil Dynamics, University of Missouri, Rolla.
Seed, R.B., S.W. Chang, et al. 1997. “Site-Dependent Seismic Response Including Recent Strong Motion
Data,” Special Session on Earthquake Geotechnical Engineering, Proc. XIV Internat ional Conference
on Soil Mechanics and Foundat ion Engineering, Balkema, Hamburg.
Seed, R.B., S.E. Dickenson, et al. 1990. “Preliminary Report on the Principal Geotechnical Aspects of the
October 17, 1989 Loma Prieta Earthquake,” Earthquake Engineering Research Center, University
of California, Berkeley, p. 137.
Seed, H.B., K. Tokimatsu et al. 1985. “The Influence of STP Procedures in Soil Liquefaction Resistance
Evaluations,” J. Geotech. Eng. ASCE, 111(12), 1425–1445.
Seed, H.B., C. Ugas, et al. 1976. “Site-Dependent Spectra for Earthquake-Resistant Design,” Bull. Seismol.
Soc. Am., 66, 221–243.
Serff, N., H.B. Seed, et al. 1976. “Earthquake-Induced Deformations of Earth Dams,” Earthquake Engi-
neering Research Center, University of California, Berkeley, p. 140.
Shabestari, K.T. and F. Yamazaki. 1998. “Attenuation Relationship of JMA Seismic Intensity Using JMA
Records,” in Proc. Tenth Japan Earthquake Engineering Symposium, Kobe, Japan.
Silver, M.L. and H.B. Seed. 1971. “Volume Changes in Sands during Cyclic Loading,” J. Soil Mech. Found.
Div. ASCE, 97(9), 1171–1182.
Somerville, P. 1998. “Emerging Art: Earthquake Ground Motion,” in Geotechnical E arthquak e Engineering
and Soil D ynamic s, Vol. III, Geotechnical Special Publication No. 75, American Society of Civil
Engineers, New York.
Somerville, P.G. and R.W. Graves. 1996. “Strong Ground Motions of the Kobe, Japan Earthquake of
January 17, 1995, and Development of a Model of Forward Rupture Directivity Applicable in
California,” in Proc. Western Regional Technical Seminar on Earthquake Engineering for Dams,
Association of State Dam Officials, Sacramento.
Somerville, P.G., C.K. Saikia, et al. 1996. “Implications of the Northridge Earthquake for Strong Ground
Motions from Thrust Faults,” Bull. Seismol. Soc. Am., 86, S115–S125.
Stark, T.D. and G. Mesri. 1992. “Undrained Shear Strength of Sands for Stability Analysis,” J. Geotech.
Eng. ASCE, 118(11), 1727–1747.
Stone, W.C., F.Y. Yokel, et al. 1987. Engineering Aspects of the September 19, 1985 Mexico Earthquak e, NBS
Building Science Series, 165, National Bureau of Science, Washington, D.C., p. 207.

© 2003 by CRC Press LLC


0068_C07_fm Page 61 Wednesday, July 31, 2002 8:20 AM

Geotechnical and Foundation Aspects 7-61

Sy, A. and R.G. Campanella. 1994. “Becker and Standard Penetration Tests (BPT-SPT) Correlations with
Consideration of Casing Friction,” Can. Geotech. J., 31, 343–356.
Tappin, D.R., T. Matsumoto et al. 1999. “Offshore Surveys Identify Sediment Slump as Likely Cause of
Devastating Papua New Guinea Tsunami 1998,” Eos, 80(30), 329.
Thiers, G.R. and H.B. Seed. 1978. Strength and S tress-Strain Char acteristics of Clays Subjected to Seismic
Loading C onditions, ASTM Special Technical Publication 450, American Society for Testing and
Materials, pp. 3–56.
Tokimatsu, K. and H.B. Seed. 1987. “Evaluation of Settlements in Sand Due to Earthquake Shaking,” J.
Geotech. Eng. ASCE, 113(8), 861–878.
Trifunac, M.D. and A.G. Brady. 1975. “On the Correlation of Seismic Intensity with Peaks of Recorded
Strong Ground Motion,” Bull. Seismol. Soc. Am., 65, 139–162.
U.S. Army Corps of Engineers. 1982. Slope Stability Manual, Department of the Army, Office of the Chief
of Engineers, Washington, D.C.
Vucetic, M. and R. Dorby. 1991. “Effect of Soil Plasticity on Cyclic Response,” J. Geotech. Eng. ASCE,
117(1), 89–107.
Whitman, R.V. and S. Liao. 1985. “Seismic Design of Retaining Walls,” U.S. Army Corps of Engineers,
Waterways Experiment Station, Vicksburg, MS.
Wieczorek, G.F. 1996. “Landslide Triggering Mechanisms,” in Landslides, Investigation and Mitigation,
A.K. Turner and R.L. Schuster, Eds., Transportation Research Board Special Report 247, National
Academy Press, Washington, D.C., pp. 76–90.
Wightman, A. 1991. “Ground Improvement by Vibrocompaction,” Geotech. News, 9(2), 39–41.
Wood, J. 1973. “Earthquake-Induced Soil Pressures on Structures,” Report EERL 73-05, California Insti-
tute of Technology, Pasadena, p. 311.
Youd, T.L. and S.N. Hoose. 1977. “Liquefaction Susceptibility and Geologic Setting,” in Proc. 6th World
Conference on Earthquak e Engineering, Prentice-Hall, Englewood Cliffs, NJ.
Youd, T.L. and I.M. Idriss. 2001. “Liquefaction Resistance of Soils: Summary Report from the 1996 NCEER
and 1998 NCEER/NSF Workshops on Evaluation of Liquefaction Resistance of Soils,” J. Geotech.
Eng. ASCE, 127(4), 297–313.
Youd, T.L. and S.K. Noble. 1997. “Magnitude Scaling Factors,” in Proc. NCEER Workshop on Evaluation
of Liquefaction Resistance of Soils, Salt Lake City, NV, Multidisciplinary Center for Earthquake
Engineering Research, Buffalo, NY.
Youd, T.L. and D.M. Perkins. 1978. “Mapping of Liquefaction-Induced Ground Failure Potential,”
J. Geotech. Eng. Div. ASCE, 104(GT4), 433–446.

Further Reading
Kramer’s book [Kramer, 1996] provides a detailed explanation of earthquake effects on geotechnical
engineering parameters, as does Finn’s article on the state of the art of geotechnical earthquake engi-
neering practice [Finn, 2001]. A series of U.S. and Japanese case studies of liquefaction effects on lifeline
performance during earthquakes contains a wealth of information on lifeline performance during
earthquakes [NCEER, 1992]. Another case study of interest is a USGS professional paper on the Marina
district of San Francisco in the 1989 Loma Prieta earthquake [O’Rourke, 1992]. An excellent review of
liquefaction remediation techniques is presented in a handbook on liquefaction remediation of reclaimed
land [PHRI, 1997].

© 2003 by CRC Press LLC

You might also like