03 - Seismic Hazards 2020

CHAPTER 3
Seismic Hazards
Downloaded from ascelibrary.org by Swinburne University of Technology on 03/27/20. Copyright ASCE. For personal use only; all rights reserved.
3.1 INTRODUCTION
Defining seismic hazards is one of the most important steps in the seismic design
and evaluation of petrochemical and other industrial facilities. Engineers who
perform seismic design are usually not responsible for the assessment and
quantification of seismic hazards; yet, owners and regulators often call on them
to interpret the results of hazard analysis and to justify their appropriateness.
Therefore, design engineers must clearly understand how specialists in geology,
seismology and geotechnical engineering develop this information.
Chapters 11, 16, 20, 21, and 22 of ASCE 7-16 cover assessment of seismic
hazards. These chapters provide specific requirements on issues ranging from
computation of ground motions for design to requirements of geologic and
geotechnical investigations and can sometimes be confusing for people not
regularly involved in assessing seismic hazards. This committee believes that a
proper understanding of the basics of seismic hazard assessment, especially
computation of ground motions, is necessary to develop an appropriate interface
among seismologists, geologists, engineers, owners, and regulators involved in the
design and evaluation process.
This chapter does not intend to instruct the engineers on how to perform
seismic hazard assessment; rather, it provides the necessary background informa-
tion on techniques used to define and quantify seismic hazards. Readers interested
in learning about the mechanics and detailed procedures of seismic hazard
analysis for developing design ground motions are referred to Seismic Hazard
and Risk Analysis, a monograph published by the Earthquake Engineering
Research Institute (McGuire 2004) and “Documentation for the 2008 Update of
the United States National Seismic Hazard Maps” by the US Geological Survey
(Petersen et al. 2008).
Hazards covered in this chapter include ground shaking; excessive ground
deformations or ground failure caused by fault rupture, liquefaction, and land-
slides; and inundation of coastal areas and resulting damage caused by large waves
produced in nearby bodies of water (tsunamis and seiches). Appendixes to this
chapter provide more detailed discussions of several specific topics.
21
Seismic Evaluation and Design of Petrochemical and Other Industrial Facilities

22 SEISMIC EVALUATION AND DESIGN OF PETROCHEMICAL
3.2 EARTHQUAKE BASICS
3.2.1 Earthquake Mechanism

Earthquakes generally result from a sudden release of built-up strain in the earth’s
crust as slip along geologic faults. Faults are classified on the basis of rupture
mechanism and the direction of movement on one side of the fault relative to the
other. The three basic types of fault rupture are strike-slip, normal, and reverse/
thrust. Figure 3-1 shows examples of fault slip and its manifestation on the ground
surface.
3.2.2 Earthquake Magnitude and Intensity

The size of an earthquake is often described in terms of its magnitude and
resulting intensity of ground shaking. The former is a measure of the overall
energy released, while the latter is a description of the effects experienced at a
particular location during the earthquake.
The magnitude is the single most commonly used descriptor of the size of an
earthquake and is typically reported as the “Richter magnitude” in the news media.
The Richter magnitude, also known as the local magnitude (ML), is one of several
magnitude scales that use the amplitude of seismic waves to compute the
earthquake magnitude. Other such magnitude scales include the surface wave
magnitude (MS) and the body wave magnitude (mb).
A limitation of such magnitude scales is that the wave amplitude does not
continue to increase proportionately with the size of the earthquake. As a result,
most magnitude scales reach an upper limit, beyond which the magnitude does
not increase commensurate with the size of the earthquake. To avoid this problem,
Hanks and Kanamori (1979) introduced the moment magnitude (MW) scale,
which is a function of the total rupture area of the fault and the rigidity of rock
rather than the amplitude of seismic waves. The moment magnitude is now the
Strike-Slip Faulting Reverse Faulting Normal Faulting

(1906 San Francisco (1980El Esnam, Algeria Earthquake) (1983 Borah Peak, Idaho
Earthquake) Source: NOAA Earthquake)
Source: USGS Source: NOAA
Photographer: G. K. Gilbert
Figure 3-1. Examples of surface fault rupture.

Source: (a) USGS, photographer G. K. Gilbert, (b) NOAA, (c) NOAA.

SEISMIC HAZARDS 23
most commonly used magnitude scale by seismologists. Appendix 3.A presents

more detailed descriptions of each of the common magnitude scales and a
comparison with the moment magnitude.
Earthquake intensity is a measure of the effects of an earthquake at a

particular location. Intensity is a function of the earthquake magnitude, distance
from the earthquake source, local soil characteristics, and other parameters (such
as the type of damage and human response). The most common measure of
intensity in the United States is the Modified Mercalli Intensity (MMI) scale,
which ranges from I (not felt) to XII (complete destruction). Appendix 3.A
provides a detailed description of the MMI scale. Although common in the United
States and many other countries, the MMI scale is not universally used; the
Japanese use the Japan Meteorological Agency (JMA) scale, which ranges from
0 to 7, and Eastern European countries use the Medvedev-Sponheuer-Karnik
(MSK) scale, with intensities that are similar to the MMI scale.
3.3 GROUND SHAKING
3.3.1 Measures of Ground Shaking for Use in Design and

Evaluation
Although the earthquake magnitude and intensity provide a fairly good estimate
of earthquake size and its general effects, they are of limited significance in
engineering design. As practiced in the industry, engineering design requires
earthquake input to be the force or displacement demands on the structure. These
demands cannot be estimated directly from either earthquake magnitude or
intensity, but can be correlated to the acceleration, velocity, or displacement of
the ground at a particular location. Of these, the ground acceleration is directly
proportional to the inertial force in the structure and is the most commonly used
parameter for design.
During an earthquake, the ground acceleration varies from one location to
another. The variation is a function of the nature and size of the earthquake, the
proximity of the site to the seismogenic fault, the wave propagation characteristics,
and the local subsurface soil conditions. Ground acceleration is typically measured
by an accelerometer, which records the amplitude of ground acceleration as a
function of time. Typically, two orthogonal horizontal and one vertical compo-
nents of ground acceleration are recorded. The acceleration time history captures
the duration, frequency content, and amplitude of earthquake motions, and it can
be integrated with respect to time to compute ground velocity and ground
displacement. The recorded acceleration time history is typically digitized over
small time increments such as 1/50 s.
Another useful representation of ground shaking is the response spectrum.
A response spectrum is the maximum response of a series of single degree of
freedom oscillators of mass (m), known period (T), stiffness (K), and constant

Output time history response of an SDOF of period T1

Acceleration
Sa1 = Maximum response

of an SDOF of period T1 and

a specified damping
Time
Spectral Acceleration
m
Sa1
K
Single Degree
of Freedom m
T = 2o
0
Oscillator (SDOF) 0 T1 Period (T)
K
Acceleration
Peak Ground
Acceleration (PGA)
Time
Input time history
Figure 3-2. Construction of response spectrum.
damping (typically 5%) plotted as a function of their period of vibration

(Figure 3-2). The spectral acceleration for a period of zero seconds represents the
response of a rigid structure that moves in phase with the ground. This is known as
the peak ground acceleration (PGA). The PGA is also the maximum acceleration of
the acceleration time history used to compute the response spectrum.
An important distinction should be made between the response spectrum of
an acceleration time history and a design response spectrum. The former describes
the response of structures of various frequencies to a specific acceleration time
history recorded at a site from a unique earthquake, while the latter represents an
accumulation of several possible time histories that is usually statistically averaged
and smoothed in some manner to define the design demand that the structure is
expected to meet in a future earthquake.
3.3.2 Factors Affecting Ground Shaking

The amplitude and duration of strong ground shaking are influenced by the
location of the site relative to the fault rupture zone, the size and mechanism of
the earthquake, and the subsurface conditions at the site. Unless modified by local
subsurface conditions, the amplitude of seismic waves typically decreases with
increasing distance from the earthquake source as they propagate through the
earth’s crust. The decay of ground motion amplitude with distance from the
earthquake source is referred to as attenuation.

SEISMIC HAZARDS 25
The common approach for predicting ground motion from an earthquake at a

particular site includes utilizing empirically derived ground motion prediction
equations (Douglas 2018; EERI 2014), which are also referred to as attenuation
relationships. A ground motion prediction equation (GMPE) is a mathematical

expression that computes the ordinates of a response spectrum as a function of the
earthquake magnitude and its distance from the site. Historically, GMPEs were
defined for generic subsurface conditions such as rock or soil; more recently, they
are defined for the average shear wave velocity for the top 30 m (approximately
100 ft) of the subsurface (referred to as Vs30). Resulting generic ground motions for
rock or soil should be modified to account for the local subsurface conditions based
on measured or estimated Vs30 for the site. Other modifications such as basin effects,
rupture directivity effects, and source mechanics are also sometime applied, as
needed. Because of the uncertainty associated with predicting ground motions and
large amount of scatter in the recorded data, a common practice is to use more than
one relationship to account for epistemic uncertainty and use a weighted average of
the results. Different relationships are available for different tectonic regimes, and
the relationship for one regime such as the Western United States should not be
applied to the Eastern United States or other parts of the world unless the areas are
believed to have similar geologic and tectonic settings. The basic assumptions
behind each GMPE must be clearly understood before their application; for
example, different relationships use different measures of distance from the
earthquake, such as closest distance to rupture or distance to the epicenter, or
they have a maximum limit on magnitude and distance for their application.
The following are several general observations regarding ground shaking
characteristics:
• The intensity of shaking is proportional to the magnitude of the earthquake. It
is also affected by the type of faulting, with thrust-type earthquakes generally
producing stronger shaking than strike-slip-type earthquakes.
• For thrust-type earthquakes, the levels of ground shaking could be as much as
20% higher on the up-thrown block or hanging wall than on the down-
thrown block or foot wall.
• Deep subduction-zone earthquakes tend to produce lower-frequency and
longer-duration ground motions compared with shallow, crustal earthquakes.
The lower-frequency-content ground motions from subduction-zone earth-
quakes tend to increase the response spectrum at longer periods.
• Subsurface conditions influence the levels of ground shaking. With everything
else held constant, ground motion at rock has stronger high-frequency
content than ground motion at soil. As a result, a rock response spectrum
has higher spectral accelerations at short periods than a soil response
spectrum, whereas the soil spectrum is higher than the rock spectrum at
longer periods (greater than about 0.5 s).
• Motion at the ground surface is influenced mainly by the dynamic char-
acteristics of the top 100 ft of the subsurface soil/rock. Softer sites with lower

Vs30 have higher spectral accelerations at longer periods of vibration (greater

than about 0.5 s). In ASCE 7, subsurface soils are classified in five different
categories from hard rock to soft soils (A through E) on the basis of Vs30 and
other geotechnical parameters. ASCE 7 defines a sixth category (F) for soils
vulnerable to potential failure or collapse under seismic loading, such as
liquefiable soils, quick and highly sensitive clays, and collapsible weakly
cemented soils. Chapter 20 of ASCE 7 describes the procedure for site
classification using geotechnical parameters or Vs30.
• Ground shaking is usually very intense in close proximity to the fault rupture
(approximately less than about 10 km), with typically a strong velocity and
displacement pulse. This is termed the near-source effect. Within the near-
source region, ground shaking in the fault-normal direction is higher than
ground shaking in the fault-parallel direction.
• Ground motion is also influenced by rupture directivity. Fault rupture
directivity increases the intensity of long-period motions (periods greater
than 0.6 s) when the rupture propagates toward the site (forward directivity)
and decreases the intensity of motions when it propagates away from the site
(backward directivity). This is particularly important for the design of tall or
slender structures with long fundamental periods.
3.4 DESIGN GROUND MOTIONS
Assessment of ground motions for design includes either a computation of a

design response spectrum or the development of acceleration time histories. Of the
two, the former is most commonly used in design. Acceleration time histories, due
to extra computational effort, are only used in special cases when a more detailed
assessment of structural response is desired or where nonlinear response must be
explicitly evaluated.
ASCE 7 provides detailed guidance on the methodology to be used for
developing the design response spectrum or acceleration time histories. Chapter 11
of ASCE 7 includes a standard approach for the computation of a design response
spectrum. As a substitution to the standard approach, the code allows the use of site-
specific analysis (Chapter 21 of ASCE 7) but restricts the resulting spectra to be no
less than 80% of the spectra obtained using the standard code approach.
The basic philosophy behind both the standard code and the site-specific
approach is to first compute ground motions representative of an extremely rare
event and then reduce them by a factor representing the inherent overstrength in a
structure to compute design level forces (FEMA P-750 2009). The code refers to
this extremely rare event as the maximum considered earthquake (MCE) and
specifies a factor of 1.5 as the inherent overstrength. The MCE, as defined in the
code, should not be confused by the older and now abandoned acronym for
maximum credible earthquake.

SEISMIC HAZARDS 27
In ASCE 7, MCE is defined as ground motions having a 2% probability of

exceedance in 50 years. This is generally true for any site within the United States
except for sites within close proximity of very active faults, such as coastal
California. In these situations, the code allows capping the probabilistically

derived 2% in 50-year ground motions by what is considered to be upper bound
motions from a single scenario earthquake. The scenario earthquake mainly
represents the maximum earthquake that a particular fault could produce
(also known as a characteristic earthquake).
ASCE 7 recognizes that the occurrence of such extreme and rare events varies
significantly across regions. Using ground motions that represent a uniform
hazard as the design basis results in very conservative design in some regions.
To provide a consistent level of design conservatism, starting with the 2010
edition, ASCE 7 has introduced risk-targeted MCE (MCER) instead of uniform
hazard MCE ground motions that were used in previous editions as the design
basis. The MCER is defined as ground motion that results in 1% probability of
collapse within a 50-year period and is obtained by iterative integration of a
site-specific ground motion hazard curve (ground motion values versus their
annual probability of exceedance) with a fragility curve for collapse (probability of
collapse as a function of ground motion).
3.4.1 Design Response Spectrum—Standard Code Approach

The MCER ground motions for the standard code approach are based on ground
motion maps for the United States developed by the US Geological Survey
(USGS). These maps provide contours of spectral acceleration at 0.2 s (also
known as short-period spectral acceleration denoted by Ss) and 1.0 s spectral
period (denoted by S1). Spectral accelerations in the code represent maxima of the
two orthogonal horizontal directions. Prior to ASCE 7-10, MCE ground motions
were based on the geometric mean of the two orthogonal horizontal directions.
The MCER design parameters (Ss and S1) for a particular site can be obtained by
entering the address or site latitude and longitude coordinates and subsurface soil
conditions (in terms of ASCE 7 site class) at the online ASCE 7 Hazard Tool
(https://asce7hazardtool.online/).
The MCER maps included in ASCE 7 are representative of a generic soft rock
site with Vs30 of 2,500 ft/s (boundary between site classes B and C) and must be
adjusted for the actual site conditions if different from the generic condition. The
first step in this process is to classify the subsurface conditions in one of the five
generic soil classes (A through E) or the sixth soil class F representing liquefiable
soils and soils that become unstable in an earthquake. The five generic site classes
range from very hard rock “site class A” to very soft soils “site class E” with the
intermediate classes B through D representing site conditions between these two
extremes.
ASCE 7 site classification is based on specific geotechnical properties of the
top 100 ft of the subsurface soil. These properties include average shear wave
velocity for the top 100 ft (Vs30), average standard penetration test (SPT) blow

counts, or average shear strength. Of these, Vs30 is considered the most reliable and
is the preferred approach for site classification. Note that in case of multiple layers
of subsurface soils, ASCE 7 includes specific formulas for computing Vs30. The
formula is based on the assumption of total travel time rather than a direct average
of velocity values for each layer. Based on the site class representative of the site
conditions, the mapped MCER ground motions for Vs30 of 2,500 ft/s are modified
by using site amplification factors Fa and Fv provided in Tables 11.4-1 and 11.4-2
of ASCE 7-16. Factors for site class B through D are based on an estimate of
average amplification of the bedrock motions for a range of generic site conditions.
Factors for site class E are about 1.3 to 1.4 times the average amplification (refer to
Chapter 11 of ASCE 7). If the site conditions are classified as site class F, ASCE 7
(Chapter 21) requires that a detailed site response analysis be performed, which
includes analytically modeling the dynamic response of the soil column above
bedrock when subjected to seismic shaking at the bedrock level. ASCE 7 also
requires the site-specific design response spectrum for site class F to not be less
than 80% of the spectrum for site class E. Detailed site-specific ground motion
procedures are also required for seismically isolated structures with S1 > 0.6g,
structures on site class E sites with Ss > 1.0g, and structures on site class D or E
sites with S1 > 0.2g. ASCE 7 lists certain conditions for exceptions to the site-
specific ground motion requirement.
MCER design parameters, after adjusting for local site response, are reduced
by a factor of 2/3 (or 1/1.5, the assumed inherent collapse margin for code-
designed structures) to obtain the short-period and 1.0 s period spectral ordinates,
SDS and SD1. These two design parameters are considered sufficient to construct
the entire design response spectrum. This is possible for two key reasons. First, the
spectral acceleration (Sa), spectral velocity (Sv), and spectral displacement (Sd) are
related to each other as a function of vibration period as shown in Equation (3-1).
Second, the general shape of a response spectrum can be idealized into three
distinct regions: the acceleration-controlled, the velocity-controlled, and the
displacement-controlled (Figure 3-3a). In each of the three regions the respective
spectral acceleration, spectral velocity, and spectral displacement are assumed to
be constant (Newmark and Hall 1982). On a log-log scale, these regions appear as
three straight lines. Response spectrum plotted in this manner is referred to as the
tripartite response spectrum (Figure 3-3b).
2
2π 2π
Sa = Sv = Sd (3-1)
T T
The transition from constant velocity to constant displacement portion of
the response spectrum is defined by a transition period (TL). ASCE 7 provides TL
contour maps for the United States.
3.4.2 Site-Specific Design Response Spectrum

Development of a site-specific design response spectrum includes consideration
of regional tectonics (faults and seismic sources, historic seismicity, earthquake

SEISMIC HAZARDS 29
A V D
Spectral Acceleration (Sa)

2.5 x PGA
S1
T
S1
S1T1
PGA T2
0 1.0 TL
Period (T)
A = Constant Acceleration Region, V = Constant Velocity Region,
D = Constant Displacement Region
Figure 3-3a. Idealized design response spectrum (spectral acceleration versus

period—linear scale).
PGA A V D
Spectral Velocity (Sv)
Period (T)
A = Constant Acceleration Region, V = Constant Velocity Region,
D = Constant Displacement Region
Figure 3-3b. Idealized design response spectrum (spectral velocity versus period—
log scale).
recurrence rates, etc.), attenuation of ground motion with distance, and the
influence of subsurface conditions at the site for which the response spectrum
is being computed. Generally speaking, site-specific design response spectra can be
computed through either a probabilistic seismic hazard analysis (PSHA) or a
deterministic seismic hazard analysis (DSHA). Local site effects in PSHA or
DSHA are either considered explicitly through a site response analysis or
implicitly by using the appropriate GMPEs. Sections 3.4.2.1 and 3.4.2.2 of this
chapter briefly describe PSHA and DSHA procedures, which are discussed in

detail in McGuire (2004). Procedures to include local site effects are discussed later
in this section and in Section 3.4.2.3.
Chapter 21 of ASCE 7 provides several rules to follow in the development of a
site-specific response spectrum. The first set of rules relates to the computation
of the MCE and the MCER. The second set of rules pertains to the development of
design response spectrum from the MCE, and finally, the third set of rules
describes the methodology for incorporating local site effects. ASCE 7 also
provides limits so that the site-specific ground motions are not significantly
different from the standard code-based values.
Per ASCE 7, the site-specific MCER can be taken as the lesser of either the
probabilistically or the deterministically derived MCER ground motions. The
probabilistic MCER spectral accelerations are computed by either applying risk
coefficients (CRS for short period and CR1 for 1.0 s period spectral ordinates) to
spectral ordinates of 5% damped uniform hazard spectrum with 2% probability of
exceedance in 50 years or through iterative integration of a site-specific hazard
curve with collapse fragility to obtain motions that result in 1% probability of
collapse. Chapter 21 of ASCE 7 refers to the two approaches as Method 1 and
Method 2, respectively. The USGS provides web-based calculators for computing
MCER spectra using both Methods 1 and 2. The deterministic MCER ground
motions are calculated as 84th percentile 5% damped ground motions from the
largest of ground motions computed for characteristic scenario earthquakes on all
known active faults in the region. ASCE 7 also imposes a lower limit on the
deterministic MCER spectrum, which is defined by a peak spectral acceleration of
1.5 g and 1.0 s spectral acceleration of 0.6g. These lower-limit values are for rock
and must be adjusted for the appropriate soil conditions at the site. For the lower-
limit deterministic MCER spectrum ASCE 7 provides different Fa and Fv factors
than those used for the standard code approach. Different Fa and Fv factors from
those in Tables 11.4-1 and 11.4-2 of ASCE 7-16 for deterministic lower bounds are
provided to adjust the shape of the spectrum, which is not accurately represented
if the factors from Tables 11.4-1 and 11.4-2 are used. If a site-specific deterministic
MCER spectrum point falls below the lower-limit spectrum, the lower-limit
spectrum point for the same period is used as the design spectrum point, unless
the probabilistic spectrum point is lower than the deterministic lower limit
spectrum point. If the probabilistic spectrum point is lower than the deterministic
spectrum point, then the probabilistic spectrum point is the design spectrum
point. This process is repeated for spectrum period points to obtain the full design
spectrum.
Once the MCER is computed, the design response spectrum is taken as 2/3 of
the MCER spectrum. The design response spectrum cannot be less than 80% of the
spectrum obtained using the regional Ss and S1 maps included in ASCE 7.
However, this restriction may not be applicable for regions that do not have
ASCE 7 Ss and S1 maps, such as locations outside the United States. In addition,
the short-period design spectral acceleration (SDS) cannot be less than 90% of the
peak spectral acceleration at any period greater than 0.2 s, and the value at 2.0 s
cannot be less than half of the 1.0 s parameter (SD1).

SEISMIC HAZARDS 31
3.4.2.1 Probabilistic Seismic Hazard Analysis

In a probabilistic seismic hazard analysis (PSHA), the probability of exceeding a
specified level of a ground motion parameter (PGA or spectral acceleration) at a
site is computed. The approach accounts for all possible earthquakes with their
specified probabilities of occurrence and typically includes simulation of hundreds
or thousands of likely earthquake scenarios that could occur anywhere within the
site region. Figure 3-4 shows the key elements of PSHA.
Computing a PSHA requires developing a mathematical model of all active
and potentially active seismic sources within the site region. This model is known
as the seismic source model. Information required for developing a source model
includes the location and the geometry of the seismic sources such as specific faults
or more generic seismotectonic provinces. Other input parameters include the
maximum magnitude potential of these sources and their seismic activity, which is
derived either from slip rate or seismicity rate. This information is used with
an appropriate GMPE to compute ground motion exceedance probabilities.
Appendix 3.A describes various elements of a PSHA.
The results from a PSHA are presented either in terms of uniform hazard
spectra for different probabilities of exceedance or a family of curves, known as
hazard curves, that show annual exceedance probabilities for each measure of
ground motion (such as PGA or spectral acceleration at specified periods of
vibration).
Most codes and criteria documents, including ASCE 7, describe the uniform
hazard spectra in terms of either return period or ground motion exceedance
probability for the design life of a structure rather than the annual probability of
exceedance. Assuming a Poisson distribution of hazard, the probability of
exceedance and return period are related as given in Equations (3-2) and (3-3).
p = 1 − e−λt (3-2)
1
λ= (3-3)
T
where
λ = Annual probability of exceedance,
p = Probability of exceedance in t years, and
T = Return period.
Commonly, building codes use 2% or 10% probability of exceedance in
50 years. Using Equation (3-2), this translates to a 2,475- or 475-year return
period, respectively. The return period as specified here is the return period
associated with a level of ground shaking and not the return period (or recurrence
interval) of an earthquake. The return period for the maximum event on a given
fault may vary from several hundred to several thousand years, depending on the
fault’s activity rate. Earthquakes can happen any time, or even more than once
during this time period.

Seismic Sources Characterization
Source 1 Area Sources

Fault Sources
Source n
N(M)
Site
Magnitude (M) Source-to-Site

Probability Density Function for M Distance (d)
(Magnitude-Recurrence Relationship)
Historic
1. Define earthquake sources and seismicity Region of interest Seismicity
2. Define magnitude-recurrence relationship for each source
3. Consider all likely magnitude-distance pairs and their probability
Select Ground Motion Prediction Compute Hazard Curve

(Attenuation) Equations
Mn Tn
Annual Frequency of
Sa (T1…..Tn)
M2 Exceedence T1
M1
Sa(T1) Sa(T2)
Distance (d) Spectral Acceleration
1. Select appropriate attenuation equation (function of source mechanism, subsurface conditions)

2. Compute spectral ordinates for each period for all likely magnitude-distance pairs
3. Compute probability of exceeding a ground motion value for each likely magnitude-distance pair
Hazard Deaggregation Develop Response Spectrum

% Contribution to Hazard
So
urc
e-to
Sa
-S it
e Dis
tan
ce
e
gn i tu d
Ma T1 T2
Period (T)
1. Plot uniform hazard spectrum representing same probability of exceedenceat each period
2. Hazard deaggregationprovides the contribution from different magnitude-distance pairs
Figure 3-4. Probabilistic seismic hazard analysis.
The hazard curve is typically plotted on a log-log scale and its shape,
especially its slope, is strongly influenced by the seismicity of the region. Typically,
a seismically active region with frequent earthquakes, such as California, has a
relatively flat slope compared with a region with infrequent large earthquakes,
such as the New Madrid region, which has a fairly steep slope. The slope of the
hazard can significantly influence design ground motions computed using ASCE 7.

SEISMIC HAZARDS 33
In regions with a relatively flat hazard curve, such as California, the two-thirds
reduction of the probabilistically derived MCE (2,475-year return period) yields
design ground motions with return periods on the order of 475 years, while for
regions such as the Central and Eastern United States, with a steep hazard curve,
this reduction leads to design ground motions with much longer return periods.
Prior to ASCE 7, design ground motions in most building codes were defined as
ground motions with a 475-year return period in all regions regardless of the
level of seismicity. With the change in the definition of design ground motions
(two-thirds MCE), there was a significant increase in these motions in low
seismicity regions. Recognizing the potential for overdesign in these regions,
starting with the 2010 edition, ASCE 7 uses MCER, which generally represents a
uniform risk of collapse. The basis for changing the design approach from uniform
hazard ground motions to risk-targeted ground motions was developed as part of
Project 07, a joint effort of the Building Seismic Safety Council, FEMA, and USGS
(Kircher et al. 2010). MCER motions are obtained by integrating a standard
structural collapse fragility curve as described in FEMA P-695 (FEMA 2009).
3.4.2.2 Deterministic Seismic Hazard Analysis

In a deterministic seismic hazard analysis (DSHA), earthquake ground motions at
a site are estimated for specific earthquake scenarios (magnitude and location
relative to the site) using appropriate attenuation relationships. Figure 3-5 shows
the key elements of a DSHA.
Deterministic analysis does not consider the likelihood of the event or the
uncertainty in its magnitude or location but does include the attenuation
uncertainty in terms of standard deviation. The uncertainty is typically accounted
for by specifying median or median plus one standard deviation ground motions
for a specific earthquake scenario. Median ground motions are typically used for
standard structures and median plus one standard deviation (84th percentile)
for essential facilities. The 84th percentile ground motions are used to define
MCER ground motions where DSHA procedures are used.
3.4.2.3 Site Response Analysis

In general, three approaches for considering local site effects are available. The first
is the direct use of a GMPE that is representative of the subsurface conditions.
Some attenuation equations distinguish only between soil and rock, while others
provide ground motion as a function of Vs30 or ASCE 7-type site classes.
The second approach for modeling local site effects is first computing rock
outcrop (surface rock) response spectra using a rock GMPE and then modifying
the rock spectrum by generic soil amplification factors such as the Fa and Fv
factors in Tables 11.4-1 and 11.4-2 of ASCE 7-16 or other published sources such
as EPRI (1993) or Stewart and Seyhan (2013).
The third approach for modeling local site effects is through a detailed
dynamic site response analysis. Such an analysis is performed when soil condition
cannot be reasonably categorized into one of the standard site conditions, or when

Seismic Sources Characterization
Area Sources
Particular source of interest

Site
Fault Sources
Source-to-Site
Distance (d)
Historic
Region of interest Seismicity
1. Define earthquake sources and seismicity

2. Select unique magnitude-distance pair of interest
Select Ground Motion Prediction (Attenuation) Equations
Mn
Sa (T1…..Tn)
M2
M1
Unique magnitude-distance pair
Distance (d)
1. Select appropriate attenuation equation (function of source mechanism, subsurface conditions)
2. Compute spectral ordinates for each period for the magnitude-distance pair of interest
Develop Response Spectrum

Sa
Period (T)
1. Plot response spectrum for a unique magnitude-distance pair
2. Develop a smoothed response spectrum, if needed
Figure 3-5. Deterministic seismic hazard analysis.
empirical site factors for the site are not available (such as for site class F).
Dynamic site response analysis can either be one-dimensional analysis that
assumes vertically propagating shear waves through the various subsurface soil
layers. Programs such as SHAKE (Schnabel et al. 1972) or DeepSoil (Hashash et al.
2016) are examples of codes used for one-dimensional equivalent linear analysis,
while programs such as FLAC (Itasca Consulting Group, 2008) and OpenSees

SEISMIC HAZARDS 35
(Regents of the University of California 2000) are examples of two-dimensional

nonlinear models.
The input to dynamic site response analysis requires identification of
subsurface soil strata; specifications of basic and nonlinear soil properties for
each layer of subsurface soil, such as unit weight and shear modulus and damping
as a function of shear strain (more advanced analysis may require additional soil
properties); and specification of input ground motion time histories. Typically, the
input time histories for dynamic site response analysis are specified as rock
outcrop (rock at ground surface) acceleration time histories that are then modified
within the program to represent bedrock (rock at depth) time histories. In cases
where the depth to bedrock is very deep and computing the bedrock time histories
is impractical, ASCE 7 allows terminating the model at depth where the soil
stiffness is at least equivalent to site class D. In such a case, site class D outcrop
time histories will be needed for site response analysis. ASCE 7 requires five
rock outcrop time histories to be either selected or simulated such that they are
representative of earthquake events that control the MCE in terms of magnitude
and distance from the fault. ASCE 7 also requires the time histories to be scaled
such that on average the response spectrum of each time history is approximately
at the level of the MCE rock spectrum over the period range of significance to
structural response. This requirement is different and less stringent than the
requirement for acceleration time histories that are directly used for structural
analysis discussed in the following section.
3.4.3 Earthquake Time Histories

In some cases, earthquake time histories are required as inputs to structural
analyses. This may be the case for nonlinear structural analysis and/or analysis of
structures of special importance and/or for analysis that considers soil-structure
interaction. Chapter 16 of ASCE 7 provides specific requirements for seismic
response history analysis that should be followed in selecting appropriate time
histories, modifying the selected time histories to meet the design requirements,
and applying the time histories in the analysis.
The approach for developing site-specific earthquake time histories should
consider the following:
• Initial selection of time histories: This includes selecting records that closely
match the site tectonic environment, controlling earthquake magnitudes and
distances, type of faulting, local site condition, response spectral character-
istics, and strong shaking duration. Both recorded time histories from past
earthquakes and synthetic time histories can be used. Multiple time histories
should be considered. Chapter 16 of ASCE 7requires a minimum of 11 time
histories for nonlinear analysis. For seismically isolated structures (ASCE 7,
Chapter 17) a minimum of seven time histories are required.
• Modification of time histories: Because the selected time histories may differ
from the design motions in terms of shaking amplitude and response spectral

ordinates, they must be modified for use in analysis. Two modification

methods can be used: amplitude scaling or spectral matching.
If amplitude scaling is used, both orthogonal components of the selected
time history are linearly scaled using the same scale factor up or down such
that the average of the maximum-direction spectra from all selected ground
motions generally matches or exceeds the target spectrum and that the
average spectrum is not less than 90% for any period within the period
range of interest of the target response spectrum. If spectral matching is used,
the time history is adjusted either in the frequency domain by varying the
amplitudes of the Fourier Amplitude Spectrum on the time domain by adding
wavelets in iterations until a satisfactory match to the target spectrum is
obtained. According to ASCE 7, the maximum-direction spectra of all-time
histories considered in spectral matching should be equal to or exceed 110%
of target spectra within the period range of interest.
The period range of interest for amplitude scaling or spectral matching
should be at least two times the largest first-mode period and not less than the
period that includes 90% of mass participation in each principal horizontal
direction. The lower-bound period should not be less than 20% of the smallest
first-mode period in each principal direction. ASCE 7 allows a reduction from
two times the upper-bound period to 1.5 times if justified by dynamic
analysis. For vertical response the lower-bound period need not be less than
0.1 s or the lowest period with significant vertical mass participation.
A uniform hazard response spectrum (in which each ordinate of the
spectrum has the same probability of exceedance) is inherently a conservative
target for scaling ground motion, because it represents ground motions for a
given hazard level and does not necessarily represent actual ground motions.
The conditional mean spectrum (CMS) provides an alternative response
spectrum for selecting ground motions (Baker 2011). A CMS is appropriate
for period-specific ground motions, and multiple CMS spectra targeting
different periods may be necessary to fully capture the structural response
in different earthquake scenarios.
3.5 GROUND FAILURE
In addition to ground shaking, seismically induced ground failures (geologic

hazards) must be evaluated. These hazards include surface fault rupture, lique-
faction, and landslides. Ground failures are important concerns for the design and
evaluation of petrochemical and other industrial facilities. An unexpected ground
failure often proves to be catastrophic, in terms of damage not only to the main
structures and equipment but also to buried systems, such as pipelines.
3.5.1 Surface Fault Rupture

Surface rupture is a direct shearing at a site during an earthquake event. It is
generally due to moderate to large earthquakes. The offsets or displacements

SEISMIC HAZARDS 37
typically have both horizontal and vertical components, and they can be as large as
several meters. A ground displacement of more than a few inches has been
observed to cause major damage to structures located on the fault. In general,
without detailed geologic investigations, the precise location of the fault is not
known with a high degree of certainty (within a few hundred feet), and because
fault displacements produce large forces and movements, the best way to limit
damage to structures is to avoid building in areas close to active faults. Empirical
relationships, such as that of Wells and Coppersmith (1994), are commonly used
to estimate the magnitudes and distribution of fault rupture displacement and
provide maximum and average fault displacements as a function of earthquake
magnitude and rupture dimensions. In California, Earthquake Fault Zone maps
(formerly known as the Alquist–Priolo maps) are available for major faults. These
maps are available online as the California Earthquake Hazards Zone Application
(“EQ Zapp”) that shows the zones of potential surface faulting and other
earthquake hazards such as liquefaction and landslide (https://maps.conservation.
ca.gov/cgs/EQZApp/app/).
3.5.2 Liquefaction
Intense shaking from an earthquake can increase the pore water pressure within
saturated soils. If the pore pressure reaches the overburden stress, the effective
stress between soil grains becomes zero and the soil liquefies. Liquefied soil is
extremely soft, has very low strength, and is easily deformed. The temporary
reduction in soil strength from liquefaction can result in foundation-bearing
failure, ground cracking, and large ground deformations. The elevated pore
pressure in liquefied soils can cause uplift of buried tanks, vaults, and pipelines.
Not all soils are susceptible to liquefaction. Assessing the susceptibility of soils
to liquefaction is the first step in liquefaction evaluation. Soil liquefaction tends to
occur primarily in saturated loose sandy or silty soils. EERI (1994) notes that
liquefaction often occurs in areas where the groundwater table is within 30 ft
(10 m) of the ground surface, and only a few instances of liquefaction have been
observed where groundwater is deeper than 60 ft (20 m). Soils must be saturated
and have little or no plasticity to be susceptible to liquefaction. For this reason,
soils may be screened for liquefaction susceptibility based on their depth relative to
the groundwater table and their Plasticity Index (PI). The National Academies of
Science, Engineering, and Medicine Report (2016) provides details about screen-
ing soils for liquefaction susceptibility.
The second step in liquefaction evaluation is an assessment of liquefaction
triggered by the anticipated intensity of ground shaking. Factors affecting
liquefaction potential include dynamic shear stresses in the soil; earthquake
magnitude or shaking duration; and soil properties including type, density,
gradation, geologic age, and depositional environment. Based on observations
from the 2010–2011 Canterbury Earthquake Sequence (CES) in New Zealand,
areas with potentially liquefiable soils can repeatedly liquefy in subsequent
earthquakes that are large enough to exceed the liquefaction threshold.

The most commonly used analytical procedures to assess the potential for
liquefaction triggering is known as the “Simplified Procedure” (Seed and Idriss
1982). The procedure uses two variables, the cyclic stress ratio (CSR) and the cyclic
resistance ratio (CRR), for this purpose. CSR characterizes the seismic demand on
the soil and is a function of depth, soil composition, effective and total overburden
stress, earthquake magnitude (M), and PGA. CRR characterizes the soil’s capacity
to resist liquefaction and is estimated using soil properties such as SPT blow count
values or cone penetrometer test (CPT) cone-tip resistance. CRR is corrected for
factors such as overburden stress, hammer energy, and fines content. The factor of
safety against liquefaction is computed as a ratio of CRR to CSR. A factor of safety
less than 1.0 indicates liquefaction is likely.
The procedures to evaluate liquefaction triggering were developed, in part,
based on empirical observations of site performances (i.e., liquefied and non-
liquefied sites) in past earthquakes. Approaches for liquefaction assessment
include Seed and Idriss (1982), Seed et al. (1983, 1985), Robertson and Wride
(1997), Youd et al. (2001), Idriss and Boulanger (2012), and Cetin et al. (2004).
Boulanger and Idriss (2014) present the latest revisions and refinements to the
procedure and include high-quality liquefaction case histories from recent earth-
quakes in New Zealand (2010–2011 CES) and Japan (2011 Mw 9.0 Tohoku).
ASCE 7 requires the potential for liquefaction to be evaluated based on the
geometric mean PGA from the maximum considered earthquake (MCEG). The
PGA from the MCEG differs from the PGA calculated using the risk-targeted
MCER. The PGA for liquefaction analysis should also reflect local site conditions,
and a site amplification factor, FPGA, which is based on the ASCE 7 site class,
should be applied to calculate the PGA. MCEG design parameters can be obtained
by entering the address or latitude and longitude coordinates of the site and
subsurface soil conditions (in terms of ASCE 7 site class) at the online ASCE 7
Hazard Tool (https://asce7hazardtool.online/).
The third step in a liquefaction evaluation is an assessment of the con-
sequences of liquefaction. Liquefaction is typically manifested in terms of bearing-
capacity failure, soil settlement, and lateral spreading. Other consequences of
liquefaction include ground motion modification, uplift of buried structures,
increased lateral loads on embedded structures, and down-drag of pile-supported
structures. The following sections discuss some of the potential consequences of
liquefaction.
3.5.2.1 Bearing Failure

The material behavior of liquefied soil is very complex, as its stiffness can vary by
many orders of magnitude throughout each cycle of loading. At very large shear
strains, the liquefied soil may approach the soil’s residual shear strength. Instabil-
ities, such as bearing-capacity failure, occur when the static shear stress in the soil
exceeds its shear strength. Therefore, the residual shear strength of liquefied soil is
an important consideration in the assessment of foundation bearing capacity.
Procedures to estimate the residual shear strength of liquefied soil are most
often backcalculated from observations of ground failure in past earthquakes.

SEISMIC HAZARDS 39
Olson and Stark (2002), Idriss and Boulanger (2008), and Robertson (2010)
develop estimates of soil residual strength based on empirical observations from
past earthquakes. Kramer and Wang (2015) present a model that is based both
on critical state soil mechanics and empirical observations. As with estimates of

ground motion intensity measures, using a weighted average of an ensemble of
residual shear strength estimates is common practice.
3.5.2.2 Settlement
Soil settlement can result from soil densification, excess pore water pressure
dissipation following liquefaction, lateral squeezing of liquefied soil under foun-
dation loads, and loss of volume from surface ejecta. Shaking from an earthquake
can cause unsaturated soils to compact and decrease in volume (volumetric
strain). The vertical accumulation of volumetric strain is reflected in post-
earthquake settlement of the ground surface. Volumetric strain also occurs as
excess pore water pressure drains from liquefied soil. The magnitude of volumetric
strain of liquefied soil is typically larger than the volumetric strain caused by
compaction of unsaturated soil. Considerations should be given to identifying the
lateral and vertical extents of liquefiable soils so that the potential for nonuniform
(differential) settlements can be evaluated. The procedures proposed by
Tokimatsu and Seed (1987), Ishihara and Yoshimine (1992), or Wu and Seed
(2004) can be used to estimate the amount of settlement as a function of
earthquake shaking and SPT blow counts.
Settlement estimates based on volumetric strains do not account for material
loss from ejecta (such as sand boils). For example, the soil may relieve excess pore
water pressure by ejecting groundwater to the surface, carrying with it fine sands
and silts, which are deposited as sand boils. Settlement caused by ejecta is difficult
to predict and depends on the depth to the water table; the ratio of nonliquefiable
crust thickness to the thickness of the underlying liquefiable soil; and existing
fractures, root holes, and structural penetrations, all of which are highly variable.
Van Ballegooy et al. (2014) discuss recent experience with settlements caused by
ejecta during the Canterbury Earthquake Sequence in New Zealand.
3.5.2.3 Lateral Spreading

Seismically induced horizontal movements can also occur on sloping or gently
sloping ground when the underlying soils liquefy. The horizontal movement of a
slope underlain by liquefied soil is termed lateral spreading. Lateral spreading can
also be initiated at free faces where lateral movement can take place. Case histories
(Zhang et al. 2004) indicate that the length of lateral spreading can extend as much
as 40 times the height of the free face. Typical procedures for lateral spread
assessment include empirical methods (Khoshnevisan et al. 2015).
Very often, the complicated behavior of liquefied soil can only be appropri-
ately evaluated using advanced, nonlinear effective stress models in programs such
as FLAC or OpenSees. These models can also assess the dynamic kinematic loads
on structures caused by laterally deforming soils. However, numerical modeling is
a significant effort and requires expertise and experience.

3.5.2.4 Liquefaction Mitigation Measures

Mitigation measures for liquefaction vary from removing liquefiable soils to
strengthening foundation systems to resist liquefaction. Table 3-1 (Seed et al.
2003) lists some of the more common mitigation methods. The selection of one or
more methods should be evaluated based on constructability, effectiveness, cost,
and other project constraints, such as environmental impacts and permitting and
project schedule. Consideration should also be given to the ability to verify the
selected methods’ effectiveness. For example, if stone columns are used to densify
liquefiable soils, SPT blow counts can be measured before and after stone column
installation so that their effectiveness can be verified. Section 12.13.9 of ASCE 7-16
provides new foundation requirements for building structures located on liquefi-
able sites.
3.5.3 Landslides
Seismically induced, or coseismic, landslides may encompass mass volume and
area that are much larger than those due to other causes. Sometimes, these
landslides involve areas of many square miles and may be located more than
100 miles from the epicenter. Most coseismic landslides have been triggered by
seismic events greater than magnitude 5.
Landslides can occur in slopes with and without the presence of liquefied
soils. Both soil strength loss and inertial loading due to shaking can contribute to
coseismic landslide deformation. In conditions where soil strength is not expected
to change significantly during shaking, pseudo-static analyses can be used to
estimate a yield acceleration, which is the acceleration that would temporarily
destabilize the slope. The pseudo-static yield acceleration can then be used in
conjunction with sliding block regression analyses, such as Bray and Travasarou
(2007) or Kim and Sitar (2004), to estimate coseismic landslide deformation.
Advanced numerical finite element or finite difference methods are necessary to
evaluate the dynamic slope responses when the soil strength degrades significantly
during shaking.
3.6 TSUNAMI AND SEICHE
Tsunamis and seiches occur regularly throughout the world, but are usually only
considered in the design and evaluation of facilities when making initial decisions
on the facility’s location. This section briefly describes these phenomena, assess-
ment methods, and possible mitigation measures. Appendix 3.B provides addi-
tional discussion.
Tsunamis are typically generated by large and sometimes distant earthquakes
associated with undersea faulting, submarine landsliding, or volcanic eruption.
Traveling through the deep ocean, a tsunami is a broad and shallow, but fast-
moving, wave that poses little danger to most vessels in deep water. When

Table 3-1. List of Methods for Mitigation of Soil Liquefaction.

General Category Mitigation Methods Notes
I. Excavation and/or Excavation and disposal of liquefiable soils

compaction Excavation and recompaction
Compaction (for new fill)
II. In situ ground densification Compaction with vibratory probes (e.g., Can be coupled with the installation of
Vibroflotation, Terraprobe, etc.) gravel columns
Dynamic consolidation (Heavy tamping)
Compaction piles/stone columns Can also provide reinforcement
Deep densification by blasting
Compaction grouting
III. Selected other types of Permeation grouting
ground treatment Jet grouting
Deep mixing
Drains: gravel drains, sand drains, prefabricated Many drain installation processes also
strip drains provide in situ densification
IV. Berms, dikes, seawalls, and Structures and/or earth structures built to provide
other edge containment edge containment and thus to prevent larger
structures/systems lateral spreading
V. Deep foundations Piles (installed by driving or vibration) Can also provide ground densification
Piers (installed by drilling or excavation)
SEISMIC HAZARDS

VI. Reinforced shallow Grade beams
foundations Reinforced mat
Well-reinforced and/or post-tensioned mat
41
“Rigid” raft
tsunamis reach the coastline, wavelengths become shortened and wave amplitudes
increase. Coastal waters may rise above normal sea level and wash inland with
great force. The succeeding outflow of water is just as destructive as the tsunami
run-up.
A seiche occurs when resonant wave oscillations form in an enclosed or semi-
enclosed body of water such as a lake, bay, or fjord. Seiches are typically caused
when strong winds and rapid changes in atmospheric pressure push water from
one end of a body of water to the other. The water then rebounds to the other side
of the enclosed water body and can oscillate back and forth for time periods
ranging from a few minutes to several days. Earthquakes, tsunamis, or severe
storm fronts may also cause seiches along ocean shelves and ocean harbors.
A tsunami or seiche may result in rapid flooding of low-lying coastal areas.
The greatest hazard results from the inflow and outflow of water, where strong
currents and forces can erode foundations and sweep away structures, vehicles,
vessels, or almost any large body in its path. Petrochemical facilities are especially
vulnerable to the rupture or movement of storage tanks from debris impact,
foundation erosion, or buoyancy, and can result in massive pollution, fires, or
explosions. The tsunami at Seward, Alaska, following the 1964 Alaska Earthquake
led to destruction of port facilities. The succession of waves spread the fire from
ruptured oil tanks across and throughout the port area. The flooding of low-lying
coastal areas from the December 26, 2004, tsunami in Southeast Asia resulted in
one of the most devastating natural disasters of the past 100 years, with more than
220,000 deaths and about 450,000 displaced homeless survivors. As a result of the
2004 tsunami, the petroleum storage and distribution facility at the deep-water
port at Kreung Raya lost half of its above-ground piping and three of 12 liquid fuel
storage tanks. None of the tanks was anchored to its foundations, and the three
that were swept away by tsunami waves were only partially full.
Tsunamis can cause more damage than strong ground shaking from an
earthquake. The 2011 tsunami in Japan overturned petrochemical storage tanks as
a result of wave impact and floating, often accompanied by the release of
significant quantities of hydrocarbons. Overall, 1,404 hazardous facilities were
damaged by the M9.0 earthquake and 1,807 by the subsequent tsunami. Produc-
tion facilities at Japan’s largest petrochemical complex, in Kashima, about 50 mi
northeast of Tokyo, were flooded with up to three ft of seawater by the 2011
tsunami, resulting in the loss of at least 1.7 million t/yr of ethylene capacity and
more than 1 million bbl/day of crude processing capacity.
A third potential cause of coastal inundation is coastal subsidence caused by
tectonic (faulting) or nontectonic (e.g., submarine landslide) effects. Permanent
coastal subsidence and submergence may occur. Predicting when, or if, such
tectonic subsidence will occur in a given area is difficult, if not impossible,
although historical earthquake data provide evidence of such occurrences. With
a nearby submarine slide, there may be no time for warning or escape.
Coastal sites along active tectonic subduction zones where dip-slip faulting is
common are most susceptible to earthquake-induced inundation. Examination of
the regional fault characteristics and earthquake history, with emphasis on

SEISMIC HAZARDS 43
submarine earthquakes and possible historical tsunamis, provides the best mea-
sure of site vulnerability. Estimates of maximum tsunami run-up may be based on
either historical occurrences or theoretical modeling.
In Japan, where tsunamis are common, several mitigation measures have been
implemented, such as sea walls and barriers to resist inundation. These structures
are sometimes unsuccessful, as large tsunamis may overtop the barriers and flood
the protected area, generating strong currents that erode the barriers and other
structures. Many people who were caught in March 11, 2011, Tōhoku Earthquake
and tsunami that resulted in 15,895 deaths and 2,539 missing thought they were
safe behind tsunami walls that had been constructed to protect against tsunamis of
much lower heights. Coastal protective structures must be carefully designed to
withstand extreme events. These protective structures can provide a false sense of
security or end up acting as powerful battering rams against the very structures
they were designed to protect.
Petrochemical facilities located along coastlines may be subject to tsunami
inundation and possible damage. Facilities should have a tsunami plan in place,
with specific actions to be taken upon notification of a tsunami. If the tsunami is
from a distant source, facilities may have hours to prepare and shut down
operations. A nearshore tsunami may only give a 5 to 10 min warning and
require a different action plan. Chapter 6 of ASCE 7 and its commentary provide
detailed guidelines for computing tsunami loads, inundation levels and mitigation
measures.
References
ASCE. 2016. Minimum design loads and associated criteria for buildings and other
structures. ASCE/SEI 7-16. Reston, VA: ASCE.
Baker, J. W. 2011. “Conditional mean spectrum: Tool for ground motion selection.”
J. Struct. Eng. 137 (3): 322–331.
Bolt, B. A. 1988. Earthquakes. New York: W. H. Freeman.
Boulanger, R. W., and I. M. Idriss. 2014. CPT and SPT based liquefaction triggering
procedures. Rep. No. UCD/CGM-14/01. Davis, CA: Univ. of California Center for
Geotechnical Modeling, Dept. of Civil and Environmental Engineering, College of
Engineering.
Bray, J. D., and T. Travasarou. 2007. “Simplified procedure for estimating earthquake-
induced deviatoric slope displacements.” J. Geotech. Geoenviron. Eng. 133 (4): 381–392.
Cetin, K. O., R. B. Seed, A. Der Kiureghian, K. Tokimatsu, L. F. Harder, R. E. Kayen, et al.
2004. “Standard penetration test-based probabilistic and deterministic assessment of
seismic soil liquefaction potential.” J. Geotech. Geoenviron. Eng. 130 (12): 1314–1340.
Cornell, C. A. 1968. “Engineering seismic risk analysis.” Bull. Seismol. Soc. Am. 58 (5):
1583–1605.
Cornell, C. A., and E. H. Vanmarcke. 1969. “The major influences of seismic risk.” In Vol. 1
of Proc., 4th World Conf. on Earthquake Engineering, Santiago, Chile, 69–83.
Der Kiureghian, A., and A. H.-S. Ang. 1977. “A fault-rupture model for seismic risk
analysis.” Bull. Seismol. Soc. Am. 67 (4): 1173–1194.
Douglas, J. 2018. Ground motion prediction equations 1964–2018. Glasgow, UK: Univ. of
Strathclyde, Dept. of Civil and Environmental Engineering.

EERI (Earthquake Engineering Research Institute). 1994. Liquefaction: Earthquake basics

brief no. 1. Oakland, CA: EERI.
EERI. 2014. “NGA-West2 Research Project.” Earthquake spectra 30 (3): 973–987.
EPRI (Electric Power Research Institute). 1993. Guidelines for determining design basis
ground motions. EPRI TR-102293. Palo Alto, CA: EPRI.
FEMA. 2009a. NEHRP recommended seismic provisions for new buildings and other
structures. FEMA P-750. Washington, DC: FEMA.
FEMA. 2009b. Quantification of building seismic performance factors. FEMA P-695.
Washington, DC: FEMA.
Gutenberg, B., and C. F. Richter. 1956. “Earthquake magnitude, intensity, energy, and
acceleration.” Bull. Seismol. Soc. Am. 46 (2): 105–145.
Hanks, T. C., and H. Kanamori. 1979. “A moment magnitude scale.” J. Geophys. Res.
84 (B5): 2348–2350.
Hashash, Y. M. A., M. I. Musgrove, J. A. Harmon, D. R. Groholski, C. A. Phillips, and
D. Park. 2016. DEEPSOIL 6.1, user manual. Urbana, IL: Board of Trustees of Univ. of
Illinois at Urbana-Champaign.
Houston, J. R., and A. W. Garcia. 1978. Type 16 flood insurance study: Tsunami predictions
for the west coast of the Continental United States. Technical Rep. No. H-78-26.
Alexandria, VA: USACE Waterways Experiment Station.
Idriss, I. M., and R. W. Boulanger. 2008. Soil liquefaction during earthquakes. Monograph
MNO-12. Oakland, CA: Earthquake Engineering Research Institute.
Idriss, I. M., and R. W. Boulanger. 2012. “Examination of SPT-based liquefaction triggering
correlations.” Earthquake Spectra 28 (3): 989–1018.
Ishihara, K., and M. Yoshimine. 1992. “Evaluation of settlements in sand deposits following
liquefaction during earthquakes.” Soils Found. 32 (1): 173–188.
Itasca Consulting Group. 2008. Fast Lagrangian analysis of continua, user’s manual, FLAC
version 6.0. Minneapolis: Itasca Consulting Group.
Kim, J., and N. Sitar. 2004. “Direct estimation of yield acceleration in slope stability
analyses.” J. Geotech. Geoenviron. Eng. 130 (1): 111–115.
Kircher, C. A., N. Luco, and A. S. Whittaker. 2010. “Project 07—Reassessment of seismic
design procedures.” In Proc., 2010 Structures Congress, Orlando, FL.
Khoshnevisan, S., H. Juang, Y. Zhou, and W. Gong. 2015. “Probabilistic assessment of
liquefaction-induced lateral spreads using CPT—Focusing on the 2010–2011 Canterbury
earthquake sequence.” Eng. Geol. 192 (Apr): 113–128.
Kramer, S. L., and C. Wang. 2015. “Empirical model for estimation of the residual strength
of liquefied soil.” J. Geotech. Geoenviron. Eng. 141 (9): 04015038.
Merz, H. A., and C. A. Cornell. 1973. “Seismic risk analysis based on a quadratic frequency
model.” Bull. Seismol. Soc. Am. 63 (6): 1999–2006.
McCulloch, D. S. 1985. “Evaluating tsunami potential.” In Evaluating earthquake hazards
in the Los Angeles region: US geological survey professional paper 1360, J. I. Ziony, ed.,
375–413. Reston, VA: USGS.
McGuire, R. K. 2004. Seismic hazard and risk analysis. Oakland, CA: Earthquake
Engineering Research Institute.
NASEM (National Academies of Science, Engineering, and Medicine). 2016. State of the
art and practice in the assessment of earthquake-induced soil liquefaction and its
consequences. Washington, DC: NASEM.
Newmark, N. W., and W. J. Hall. 1982. Earthquake spectra and design. Oakland, CA:
Earthquake Engineering Research Institute.

SEISMIC HAZARDS 45
Olson, S. M., and T. D. Stark. 2002. “Liquefied strength ratio from liquefaction flow failure
case histories.” Can. Geotech. J. 39 (3): 629–647.
Petersen, M. D., A. D. Frankel, S. C. Harmsen, C. S. Mueller, K. M. Haller, R. L. Wheeler,
et al. 2008. Documentation for the 2008 update of the United States National Seismic
Hazard Maps. USGS Open-File Rep. No. 2008-1128. Reston, VA: USGS.
Power, M., B. Chiou, N. Abrahamson, Y. Bozorgnia, T. Shantz, and C. Roblee. 2008.
“An overview of the NGA project.” Earthquake Spectra 24 (1): 3–21.
Regents of the University of California. 2000. Open system for earthquake engineering
simulation (OpenSees). Berkeley, CA: Pacific Earthquake Engineering Research Center.
Reiter, L. 1990. Earthquake hazard analysis issues and insights. New York: Columbia
University Press.
Richter, C. F. 1958. Elementary seismology. San Francisco: W. H. Freeman.
Robertson, P. K. 2010. “Evaluation of flow liquefaction and liquefied strength using the cone
penetration test.” J. Geotech. Geoenviron. Eng. 136 (Jun): 842–853.
Robertson, P. K., and C. E. Wride. 1997. “Cyclic liquefaction and its evaluation based on
SPTand CPT.” In Proc., Seismic Short Course on Evaluation and Mitigation of Earth-
quake Induced Liquefaction Hazards, NCEER Workshop, San Francisco.
Schnabel, P. N., J. Lysmer, and H. B. Seed. 1972. SHAKE, a computer program for
earthquake response analysis of horizontally layered sites. EERC Rep. No. 72-12. Berkeley,
CA: Univ. of California Earthquake Engineering Research Center.
Seed, H. B., and I. M. Idriss. 1982. Vol. 5 of Ground motions and soil liquefaction during
earthquakes. Engineering Monograph. Oakland, CA: Earthquake Engineering Research
Institute.
Seed, H. B., I. M. Idriss, and I. Arango. 1983. “Evaluation of liquefaction potential using field
performance data.” J. Geotech. Eng. 109 (3): 458–482.
Seed, H. B., K. Tokimatsu, L. F. Harder, and R. M. Chung. 1985. The influence of SPT
procedures in soil liquefaction resistance evaluations. Rep. No. UCB/EERC-84/15.
Berkeley, CA: Univ. of California Earthquake Engineering Research Center.
Seed, R. B., K. O. Cetin, R. E. S. Moss, A. M. Kammerer, J. Wu, J. M. Pestana, et al. 2003.
“Recent advances in soil liquefaction engineering: A unified and consistent framework.”
In Proc., 26th Annual ASCE Los Angeles Geotechnical Spring Seminar, Keynote Presen-
tation. Long Beach, CA.
Stewart, J. P., and E. Seyhan. 2013. Semi-empirical nonlinear site amplification and its
application in NEHRP site factors. PEER Rep. No. 2013. Berkeley, CA: Univ. of California
Pacific Earthquake Engineering Research Center.
Tokimatsu, K., and H. B. Seed. 1987. “Evaluation of settlements in sand due to earthquake
shaking.” J. Geotech. Eng. 113 (8): 861–878.
UBC (Uniform Building Code). 1997. International conference of building officials.
Whittier, CA: UBC.
van Ballegooy, S., S. C. Cox, R. Agnihotri, T. Reynolds, C. Thurlow, H. K. Rutter, et al. 2013.
Median water table elevation in Christchurch and surrounding area after the
4 September 2010 Darfield Earthquake Version 1. GNS Science Rep. No. 2013/01. Lower
Hutt, New Zealand: GNS Science.
Wells, D. L., and K. J. Coppersmith. 1994. “New empirical relationships among magnitude,
rupture length, rupture width, rupture area, and surface displacement.” Bull. Seismol. Soc.
Am. 84 (4): 974–1002.
Wu, J., and R. B. Seed. 2004. “Estimation of liquefaction-induced ground settlement
(case studies).” In Proc., 5th Int. Conf. on Case Histories in Geotechnical Engineering,
Paper No. 3.09. New York.

Youd, T. L., I. M. Idriss, R. D. Andrus, I. Arango, G. Castro, J. T. Christian, et al. 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.
Geoenviron. Eng. 127 (10): 817–833.

Youngs, R. R., and K. J. Coppersmith. 1985. “Implications of fault slip rates and earthquake
recurrence models to probabilistic seismic hazard estimates.” Bull. Seismol. Soc. Am.
75 (4): 939–964.
Zhang, G., P. K. Robertson, and R. W. I. Brachman. 2004. “Estimating liquefaction-induced
lateral displacements using the standard penetration test or cone penetration test.”
J. Geotech. Geoenviron. Eng. 130 (8): 861–871.
APPENDIX 3.A GROUND SHAKING
3.A.1 INTRODUCTION
The most widely used seismic input parameter in the seismic design of structures
is a measure of the ground shaking, usually given in terms of ground acceleration.
It is often the only parameter used for specification of seismic design.
Section 3.2 provides background information on common terminology and
fundamentals related to the evaluation of ground-shaking hazards. This Appendix
provides additional discussion on specific areas that are often misunderstood,
including terminology and interpretation and the application of its results. This
section aims to explain common ground motion descriptions and their relevance
to engineering applications.
3.A.2 EARTHQUAKE BASICS
Earthquakes are generally caused by the sudden release of built-up elastic strain in
the earth’s crust and originate as slip along geologic faults. The total energy of an
earthquake is released along the entire length of the ruptured zone, propagating
from the source and traveling through the earth in the form of seismic waves. The
point in the earth’s crust where the rupture is initiated is known as the focus or
the hypocenter of the earthquake. The point on the surface of earth directly above
the hypocenter is known as the epicenter of the earthquake.
The theory of plate tectonics can explain the build-up of elastic strain in the
earth’s crust, positing that the outer 70–150 km of the earth’s crust comprises
approximately 12 major plates that are slowly moving relative to each other. Most
of the seismically active areas are located along the plate boundaries. The relative
motion between the plates results in either plates grinding past each other, a plate
subducting beneath another, or several plates converging and crushing smaller
plates. There are also divergent plate boundaries. Mountains and deep sea trenches

SEISMIC HAZARDS 47
are formed where plates converge and subduct beneath one another. Mid-ocean
ridges are formed at divergent plate boundaries. Earthquakes usually initiate at
shallow depths at plate boundaries where plates slide past each other. Deep focus
earthquakes usually occur along subduction plates. Earthquakes within a plate can
also result for other reasons such as increased compressional stress. Other causes
of earthquakes include nuclear explosions or large artificial reservoirs that change
the local state of stress within the earth’s crust. Also, volcanic activity can cause
earthquakes at both subducting and diverging plate boundaries.
3.A.3 MAGNITUDE AND INTENSITY
The amount of energy released in an earthquake and its size characterize the
magnitude of the earthquake. Some of the most commonly used magnitude scales
are Richter or local magnitude (ML), surface wave magnitude (Ms), body wave
magnitude (mb), and moment magnitude (Mw).
Gutenberg and Richter (1956) show that the amount of energy released in an
earthquake can be related to the magnitude by the following relationship:
log E = 1.5M S þ 11.8 (3.A-1)

where E is the energy in ergs. A magnitude increase by one unit releases 31.6 times
more energy.
Richter (1958) defined the magnitude of a local earthquake (ML) as the
“logarithm to base ten of the maximum seismic-wave amplitude (in thousandth of
a millimeter) recorded on a standard seismograph at a distance of 100 kilometers
from the earthquake epicenter” (Bolt 1988). Seismograms from deep focus
earthquakes differ significantly from those for shallow focus earthquakes, even
though the total amount of energy released may be the same. This results in
different magnitude estimates for earthquakes that release the same amount of
energy. This is a limitation of ML.
The magnitude of an earthquake is also determined by the amplitude of the
compressional or P wave. The focal depth of an earthquake does not affect the P
wave’s amplitude. The magnitude estimated by measuring the P wave amplitude is
called the P wave or body wave magnitude (mb).
Magnitude calculated by measuring the amplitude of long-period surface
waves (with periods near 20 s) is called the surface wave magnitude (Ms). Surface
wave magnitude is calculated only for shallow focus earthquakes that give rise to
surface waves.
For small or moderate size earthquakes, the amplitude of seismic waves
measured by typical or standard seismographs increases as the size of the
earthquake increases. This trend, however, does not continue for large or very
large earthquakes for which the amplitude of seismic waves, whose wave lengths
are much smaller than the earthquake source, do not increase proportionally to

the size of the earthquake. The magnitude estimated by the amplitude of seismic
waves, therefore, does not continue to increase at the same rate as the size of the
earthquake. Beyond a certain limit the magnitudes calculated in this way tend to
remain constant for different size earthquakes. In other words, the magnitude
scale saturates, resulting in no further increase in estimated magnitude with the
increasing size of an earthquake.
Because of the limitations in the previously defined magnitude scales, Hanks
and Kanamori (1979) introduce a magnitude scale based upon the amount of
energy released, which is termed the moment magnitude scale (Mw). Moment
magnitude is defined by the following relationship:
2
M w = log M o − 10.7 (3.A-2)
3
where
M o = μ AD (3.A-3)
The seismic moment Mo is defined as a product of the rock’s modulus of

rigidity (μ), area of rupture (A), and average fault displacement (D) in a seismic
event.
MW, therefore, depends on the size of rupture, unlike other magnitude scales
that depend upon the amplitude of the seismic waves. Larger earthquakes have
larger seismic moment (product of rupture area and average fault displacement).
The moment magnitude scale, therefore, does not saturate with the size of the
earthquake and also provides the ability to distinguish between a large and a great
earthquake. For example, the surface wave magnitude for both the 1906
San Francisco Earthquake and the 1960 Chile Earthquake was estimated as
8.3, although the rupture area of the 1960 earthquake was about 35 times greater
than that of the 1906 earthquake. The moment magnitude for the two earthquakes
has been computed to be approximately 8.0 and 9.5, respectively (Reiter 1990).
Figure 3.A-1 shows an approximate relation between the various magnitude
scales. According to this figure, except for MS, all other magnitude scales are
similar to MW up to the point where they start to saturate.
Another descriptor of the size of an earthquake is the intensity at a given site.
The intensity at any location is described by a qualitative scale that uses eyewitness
accounts of fault motion and damage assessments to describe the amount of
movement felt at that location. One commonly used scale is the Modified Mercalli
intensity (MMI). As Table 3.A-1 shows, the MMI scale consists of 12 intensity
levels. Whereas intensity I means the earthquake is practically not felt, intensity
XII indicates almost total destruction. Intensity XII is rare except in very large
earthquakes. Intensity X can occur in moderate to large earthquakes, especially in
areas close to the rupture zone. Engineered structures can be damaged in areas
experiencing intensities in the range of VIII–X.

SEISMIC HAZARDS 49
9
MS
MJMA
8
mS
ML
7
mb
Magnitude
4
S
L
M
2
2 3 4 5 6 7 8 9 10
Moment Magnitude, Mw
Figure 3.A-1. Relation between Moment Magnitude and other magnitude scales:
ML (local), Ms (surface wave), mb (short-period body wave), mg (long-period body
wave), and MJMA.
Source: Japan Meteorological Agency.
3.A.4 SITE-SPECIFIC RESPONSE SPECTRA
Two distinct approaches are generally used for site-specific estimates for devel-
oping seismic design criteria for engineered projects: the deterministic procedure
and the probabilistic procedure. Descriptively, the two procedures would at first
appear to be quite different with a deterministic approach offering the advantage
of appearing easiest to follow. Properly applied, either procedure can lead to
satisfactory seismic design.
In areas of low seismic hazard, the additional design and material costs that
ensue when unsure but conservative decisions are made may not be excessive.
When the seismic exposure is high, the cost associated with additional conserva-
tism can increase significantly. This increase needs to be balanced with the
likelihood that the damaging event may or may not occur during the life of the
structure. Large damaging earthquakes are infrequent events that may simulta-
neously affect a large number of structures. The extent of damage during an
earthquake can be related to the level of motions generated by the earthquake. The
level of earthquake motion against which design or evaluation criteria should be
developed requires subjective judgment based on experience and observation.
Probabilistic methods were developed as a result of a desire to quantify some of

Table 3.A-1. Modified Mercalli Intensity Scale.
I. Not felt. Marginal and long-period effects of large earthquakes.

Felt by persons at rest, on upper floors, or favorably placed.
II.
III. Felt indoors. Hanging objects swing. Vibration like passing of light
trucks. Duration estimated. May not be recognized as an
earthquake.
IV. Hanging objects swing. Vibration like passing of heavy trucks, or
sensation of a jolt like a ball striking the walls. Standing motor cars
rock. Windows, dishes, doors rattle. Glasses clink. Crockery clashes.
In the upper range of IV wooden walls and frames creak.
V. Felt outdoors; direction estimated. Sleepers wakened. Liquids
disturbed, some spilled. Small unstable objects displaced or upset.
Doors swing, close, open. Shutters, pictures move. Pendulum
clocks stop, start, change rate.
VI. Felt by all. Many frightened and run outdoors. Persons walk
unsteadily. Windows, dishes, glassware broken. Knickknacks,
books, etc., off shelves. Pictures off walls. Furniture moved or
overturned. Weak plaster and masonry D cracked. Small bells ring
(church, school). Trees, bushes shaken (visible, or heard to rustle).
VII. Difficult to stand. Noticed by drivers of motor cars. Hanging objects
quiver. Furniture broken. Damage to masonry D, including cracks.
Weak chimneys broken at roof line. Fall of plaster, loose bricks,
stones, tiles, cornices (also unbraced parapets and architectural
ornaments). Some cracks in masonry C. Waves on ponds; water
turbid with mud. Small slides and caving in along sand or gravel
banks. Large bells ring. Concrete irrigation ditches damaged.
VIII. Steering of motor cars affected. Damage to masonry C; partial
collapse. Some damage to masonry B; none to masonry A. Fall of
stucco and some masonry walls. Twisting, fall of chimneys, factory
stacks, monuments, towers, elevated tanks. Frame houses moved
on foundations if not bolted down; loose panel walls thrown out.
Decayed piling broken off. Branches broken from trees. Changes in
flow or temperature of springs and wells. Cracks in wet ground and
on steep slopes.
IX. General panic. Masonry D destroyed; masonry B seriously damaged.
(General damage to foundations.) Frame structures, if not bolted,
shifted off foundations. Frames racked. Serious damage to
reservoirs. Underground pipes broken. Conspicuous cracks in
ground. In alluviated areas, sand and mud ejected, earthquake
fountains, sand craters.
(Continued)

SEISMIC HAZARDS 51
Table 3A-1. Modified Mercalli Intensity Scale. (Continued)
X. Most masonry and frame structures destroyed with their foundations.

Some well-built wooden structures and bridges destroyed. Serious

damage to dams, dikes, embankments. Large landslides. Water
thrown on banks to canals, rivers, lakes, etc. Sand and mud shifted
horizontally on beaches and flat land. Rails bent slightly.
XI. Rails bent greatly. Underground pipelines completely out of service.
XII. Damage nearly total. Large rock masses displaced. Lines of sight and
level distorted. Objects thrown into the air.
Notes: To avoid ambiguity, the quality of masonry, brick, or other material is specified by the
following lettering system. (This has no connection with the conventional classes A, B, and C
construction.)
Masonry A. Good workmanship, mortar, and design; reinforced, especially laterally, and bound
together by using steel, concrete, etc.; designed to resist lateral forces.
Masonry B. Good workmanship and mortar; reinforced, but not designed to resist lateral forces.
Masonry C. Ordinary workmanship and mortar; no extreme weaknesses, like failing to tie in at
corners, but neither reinforced nor designed to resist horizontal forces.
Masonry D. Weak materials, such as adobe; poor mortar; low standards of workmanship; weak
horizontally.
this judgment understanding and to allow other practitioners to use it in a

repeatable way.
The following sections discuss both the deterministic and probabilistic
seismic hazard approaches.
Deterministic Seismic Hazard Analysis

In a deterministic approach, ground motions at a site are estimated by considering
a single event of a specified magnitude and distance from the site. To perform a
deterministic analysis, the following data are required:
(a) Definition of an earthquake source (e.g., a known geologic fault) and its
location relative to the site;
(b) Definition of a design earthquake that the source is capable of producing;
and
(c) A relationship that describes the attenuation of the ground motion
parameter of interest, for example, peak ground acceleration or response
spectral ordinates for a specific natural period or frequency of vibration.
An earthquake source in most cases is a known active or potentially active
geologic fault. A site may have several faults nearby. All these sources must be
identified. Based on the length and characteristics of the fault, a maximum
magnitude potential of each source is specified. Estimation of maximum magni-
tude potential of geologic sources can be obtained from published sources or
qualified professionals. The maximum magnitude potential of a source is deter-
mined from empirical relationships between magnitude and rupture length or

through historical knowledge of past earthquakes on the particular source.

Geologists estimate the maximum magnitude potential based on the character-
istics of the source and factors such as historical activity and source dimensions.
Several relationships exist that relate source dimensions to its maximum magni-
tude potential. This information is then used as input to an attenuation relation-
ship, which relates the ground motion parameter of interest to magnitude and
distance.
Attenuation of ground motion is known to vary significantly during any
earthquake. While the variation is widely known, the quantification of the reasons
for the variation is not fully understood. Examples of possible causes may be
different rupture mechanisms in the epicenter area, directionality of the radiation
pattern of the motion, and regional geological variations. Researchers using
statistical analysis of large ground motion data sets have shown that the scatter
of ground motion values about a mean attenuation relationship can be represented
by a log-normal distribution. Attenuation relationships are empirical relationships
developed by using regression analysis techniques on ground motion data
recorded during earthquakes. Several different attenuation relationships are
available in published sources and are a function of earthquake magnitude, its
distance, local geology, and tectonic environment (subduction zones of the Pacific
Northwest, eastern United States, and western United States). Power et al. (2008)
summarize the recent attenuation relationships.
Because attenuation relationships are empirical relationships derived using
statistical analysis techniques on recorded data, they do not fit each and every data
point exactly. The actual recorded data spreads around the median attenuation
relationship. This spread is defined by the standard deviation. The attenuation
equation with no standard deviation means that 50% of the recorded values are
above and 50% are below the predicted value. An attenuation relationship with
one standard deviation means that 84% of the recorded values are below the
predicted values; similarly, attenuation relationships with two standard deviations
imply that 98% of the recorded data are below the predicted value.
The definitions of earthquake magnitude and source-to-site distance some-
times vary from one attenuation relationship to another. The magnitude scale and
source-to-site distance used in the analysis must be consistent with the proper
definition for the particular attenuation relationship used. If the site has several
sources in close proximity, an envelope of values of response spectra developed for
each source can be used.
Probabilistic Seismic Hazard Analysis

Probabilistic seismic hazard analyses (PSHA) require additional seismic criteria
beyond the selected magnitude, minimum distance, and the attenuation equations
used for the deterministic procedure defined in the previous section. The primary
additional requirements are the dimensions of each source region and the means
to interpret the probability of occurrence within each source.
Probabilistic seismic hazard analysis procedures were first developed and
used by Cornell (1968) and Cornell and Vanmarcke (1969) and further extended

SEISMIC HAZARDS 53
by Merz and Cornell (1973) and Der Kiureghian and Ang (1977). This method
involves careful consideration and incorporation of the geology, history, and
tectonics of the site region into a seismotectonic source model. A probabilistic
seismic hazard study requires the acceptance of several assumptions regarding

earthquakes.
The first step in PSHA requires that the seismic history and geology are
sufficiently well known to permit an estimate of the region’s seismic activity. This
is defined as the region’s seismicity. This information is then used to develop a
seismic source model. Potential sources can be defined as either point sources, line
sources, or area sources (Der Kiureghian and Ang 1977). Considerations may be
needed to account for the random occurrence of small or moderate sized earth-
quakes (background seismicity).
The probability distribution of earthquakes to obtain the relative occurrence
frequency distribution between large and small earthquakes is then obtained.
Estimation of recurrence rate is a key parameter in PSHA procedures. Two
approaches for estimating the recurrence rate are generally used. One is based on
the historical seismicity records, and the other is a geologic slip rate model. The
former is based on the recorded seismic history of the region, whereas the latter
uses geologic information to define strain rate accumulation using long-term slip-
rate data for a fault.
Recurrence rates can be expressed in the form of the familiar Richter’s law of
magnitudes (Richter 1958), expressed as
log N = a − bM (3.A-4)
where N is the number of events of magnitude M or greater for the time period
under consideration, and a and b are constants that depend on the seismicity of
the region.
Other recurrence models, such as the characteristic earthquake model
(Youngs and Coppersmith 1985), may also be used.
For PSHA, the minimum value of 5.0 is usually considered for M, because
earthquakes of sizes less than 5.0 are generally not damaging to engineered
structures. Due to the integrative nature of PSHA, including earthquakes of
magnitude less than 5.0 would tend to increase the ground motion hazard for the
peak ground acceleration and short periods of the response spectrum for small
return periods because of the relatively high frequency of occurrence of small
earthquakes, which are not necessarily damaging to structures.
The PSHA procedures combine the probability of occurrence of an earth-
quake, the probability of it being a specific size, and the attenuation of its motion
to the site, to obtain the probability of exceedance of a specified ground motion
level. Combination of these individual probabilities for the different source
zones at each motion level of interest then provides the total annual probabilities
of exceedance. Ground motions corresponding to a particular probability level
can then be obtained by interpolation. Figure 3-4 of Chapter 3 shows the
different elements of a PSHA graphically. Several computer programs are available
for PSHA.

The choice of probability levels for ground motion usually depends upon the
design criteria for each individual project and is established based on several
parameters such as level of risk and importance of the facility. Various probability
levels commonly used are 50% probability of exceedance in 50 years, 10%

probability in 50 years, and 10% probability of exceedance in 100 years. Assuming
a Poisson distribution, these probability levels correspond to an equivalent return
period of 72-, 475-, and 950-year return periods, respectively, or annual exceed-
ance probabilities of 0.014, 0.002, and 0.001. The relation between return period,
T, and probability of exceedance for a Poisson distribution is given as
−t
T= (3.A-5)
lnð1 − pÞ
where p is the probability of exceedance in t years. In this equation, the probability

of exceedance, p, is expressed as a decimal (e.g., 0.10 stands for 10%).
A response spectrum developed from a PSHA is referred to as a uniform
hazard spectrum, because each ordinate has the same associated probability of
exceedance (a constant level of risk). Because the response spectrum from a PSHA
analysis is calculated by a sum of individual probabilities of occurrence of several
earthquakes of different magnitudes and distances within the entire area of study,
it cannot be related to an earthquake of specified magnitude and distance from the
source.
The level of ground motion corresponding to a 2% probability of exceedance
in 50 years is the basis of the maximum considered earthquake (MCE) in ASCE 7.
In the Unified Building Code (UBC 1997), the design basis is the ground motion
with a 10% probability of exceedance in 50 years.
Uncertainty associated with the selection of input parameters must be
included in a PSHA. Uncertainty could be in the definition of magnitude-
recurrence relationship, or in the definition of slip rates, or the maximum
magnitude potential, or geographical location of a source. Furthermore, uncer-
tainty also exists in defining the ground motions. The uncertainty distribution on
the attenuation model indicates that no matter how accurately we know the
magnitude and distance of a postulated earthquake, some uncertainty will still
remain in predicting what the ground motion will be (Reiter 1990). The
uncertainty associated with an attenuation relationship (standard deviation) is
incorporated in the seismic hazard during the integration process to calculate the
probability of occurrence of a specified ground motion level.
Sometimes a tendency exists to make conservative estimates of input
parameters to be used as an input to a PSHA. This is not consistent with
the philosophy of PSHA. Input parameters should be selected based on “best
estimate” values. The uncertainty on the best estimate can be incorporated in the
analysis using a logic tree approach, where multiple estimates are specified and
weighted.

SEISMIC HAZARDS 55
APPENDIX 3.B EARTHQUAKE-RELATED COASTAL INUNDATION
3.B.1 INTRODUCTION
Section 3.6 describes briefly the phenomena associated with tsunamis and seiches.
This appendix discusses in more detail several issues associated with earthquake-
induced coastal inundation.
3.B.2 EARTHQUAKE-INDUCED COASTAL INUNDATION HAZARDS
One or more of the following earthquake-related inundation hazards may affect a

given site in a coastal area:
(a) Coastal subsidence caused by tectonic (faulting) or nontectonic (e.g.,
submarine landslide) effects;
(b) Tsunamis from either large (MW > 6.5) local earthquakes or distant great
(MW > 7.7) earthquakes; and
(c) Seiches in semi-enclosed bays, estuaries, lakes, or reservoirs caused by
moderate (MW > 5.0) local (submarine) earthquakes or regional large
earthquakes.
Each of these hazards is described in detail below.
3.B.2.1 Coastal Subsidence

In areas of dip-slip faulting (normal, thrust, or oblique-slip faulting) that involves
a vertical component of movement, coseismic subsidence along coastal faults
could result in permanent submergence of the coastal area. Based on global
experiences such as the 1964 Alaska (MW = 8.4); the 1992 Cape Mendocino,
California (MW = 7.2); and the 2004 Banda Aceh, Indonesia earthquakes, tectonic
subsidence of more than 6 ft (2 m) could occur. Furthermore, earthquakes on
blind faults, such as the 1994 Northridge California Earthquake (MW = 6.7), may
create significant surface uplift or subsidence without attendant surface fault
rupture.
Permanent coastal subsidence may also occur as a result of earthquake-
induced seafloor slumping or landslides. During the 1964 Alaska Earthquake, the
Seward port area, which included petrochemical facilities, progressively slid under
the bay as shaking continued. This subsidence was followed by withdrawal of the
sea and subsequent inundation of the coast by a wave 30 ft high. In addition to this
slide-induced wave, tsunami waves some 30 to 40 ft high, generated by the tectonic
deformation, arrived later and compounded the destruction of Seward. Similar
effects on a lesser scale were experienced at Moss Landing on Monterey Bay
during the 1989 Loma Prieta, California Earthquake (MW = 7.1) even though the

epicenter was located onshore in the Santa Cruz Mountains. At Moss Landing,
subsidence and related liquefaction created large fissures, generated a small
tsunami in Monterey Bay, and caused substantial damage to the Moss Landing
Marine Facilities. Thus, sites located where landslides or unstable slopes are
common, both on land and offshore, should assume that submarine slumping and
coastal inundation could occur during large earthquakes in the region.
For disaster planning and emergency response purposes, significant coastal
subsidence followed by tsunami run-up should be considered likely to exceed run-
up predicted based on presubsidence topography alone. Combined subsidence
and tsunami effects are considered the likely cause of the destruction to the city of
Port Royal on the Island of Jamaica in 1692 and could occur in many seismically
active areas of the world. Thus, maps of potential inundation should consider
these combined effects where coastal subsidence may be expected.
3.B.2.2 Tsunami (Seismic Sea Wave)

Tsunamis are long-period (T from 5 to 60 min) surface gravity waves, with
wavelengths that may exceed 60 mi (100 km). Tsunamis are typically generated by
large submarine earthquakes that displace the seafloor over large areas. The
destructive effects of tsunami waves may be localized, occurring along the coasts
situated close to the tsunami origin, or the waves may travel with little attenuation
across entire oceans and affect coasts thousands of miles away.
Earthquake-induced seafloor displacement may occur with or without sea-
floor fault rupture (as on a blind fault) and could generate a potentially destructive
tsunami. Even strike-slip faults, although less likely to generate significant vertical
seafloor displacement than dip-slip faults, may cause substantial seafloor uplift or
subsidence in places where these faults bend and curve. Furthermore, large
earthquakes centered on faults near the coast but on land (e.g., the 1992 Cape
Mendocino, California Earthquake) may also cause local tsunamis because of the
broad regional tectonic uplift (or subsidence) they create.
In addition to primary tectonic deformation associated with the earthquake,
large-scale slumping or landslides under the ocean or other large water bodies may
generate significant local tsunamis. A large rockslide in Lituya Bay, Alaska,
triggered by an earthquake in 1957 generated a tsunami that surged 1,700 m up
the side of the fjord. Volcanic explosions, such as that of Krakatoa in 1883, may
also generate destructive tsunamis that inundate surrounding coasts. Coastal areas
in active tectonic regions are more likely to be susceptible to tsunami attack.
However, large tsunamis may travel across entire ocean basins and affect areas
with little or no local tectonic activity.
Around the Pacific Ocean basin, often referred to as the “Ring of Fire,” most of
the world’s great earthquakes, volcanic eruptions, and destructive tsunamis occur.
Earthquakes in this region have large dip-slip (vertical) movements that displace the
seafloor rapidly, initiating the sea surface disturbance that becomes a tsunami.
Places such as Japan, Alaska, Hawaii, and western South America are particularly
vulnerable to these tsunamis, both because they have many large earthquakes
(and volcanoes) locally situated and because their geographic position and coastal

SEISMIC HAZARDS 57
configuration expose them to distantly generated tsunamis. In the contiguous

United States, recent evidence uncovered along the coasts of Washington, Oregon,
and northern California show that large local tsunamis have been generated by great
earthquakes in the Cascadia subduction zone. The Atlantic coast has also suffered
destructive tsunamis such as the one generated by the great Lisbon Earthquake in
1755, but such events are rare and generally less severe on the US side of the Atlantic
Ocean. Whereas the Gulf of Mexico has been relatively free of destructive tsunamis,
in the Caribbean, active tectonics along the Puerto Rico trench and Antilles arc have
generated notable tsunamis. Active tectonism with large earthquakes, volcanoes,
and tsunamis also occurs around the Indian Ocean.
Propagation of tsunamis often results in dispersion of the initial tsunami wave
form so that several significant waves may strike the coast. The first wave to arrive
may be smaller than subsequent waves, and persons in affected areas must realize
that a drawdown following one wave may soon be followed by larger, more
destructive, wave arrivals. The run-up speed of such waves (1 to 20 m/s or 20
to 40 mi/h) would outpace even the fastest human runners. In bays or other partially
enclosed basins, tsunamis may also set up a seiche (see Section 3.B.2.3), which can
amplify wave height and prolong the tsunami’s duration. For example, following
the 1964 Alaska Earthquake (MS = 8.4), the tsunami recorded at Santa Monica,
California, measured up to 2 m (6.5 ft) high and oscillations of half this amount
continued to occur 17 to19 hours after the arrival of the first wave. Consequently,
destructive waves from a tsunami may persist for a long time, and vulnerable areas
should remain evacuated until authorities have broadcasted an “All Clear” message.
3.B.2.3 Seiche
A seiche occurs when resonant wave oscillations form in an enclosed or semi-
enclosed body of water such as a lake, bay, or fjord. Seiches may be triggered by
moderate or larger local submarine earthquakes, and sometimes by large, distant
earthquakes. Tectonic deformation of Hebgen Lake during the 1959 earthquake
initiated seiche oscillations of up to 8 ft over the new lake level (which changed as a
result of the deformation). Such oscillations have also been triggered by meteo-
rological disturbances. For example, the passage of a storm front across Lake
Michigan tends to pile water on the eastern shore as the front advances. After the
front passes, the water flows back westward, setting up the seiche. The initial storm
wave setup is similar to storm surges observed along coasts affected by hurricanes
and other tropical storms, but the seiche results from the resonant oscillation of
the wave due to the enclosed character of the water body. Seiche oscillations may
persist for several hours.
3.B.3 SITE VULNERABILITY ASSESSMENT
Coastal sites in active tectonic areas where dip-slip faulting is common are most
susceptible to earthquake-induced inundation. Examination of the regional fault

characteristics and earthquake history, with emphasis on submarine earthquakes

and possible historical tsunamis, provides the best measure of site vulnerability.
Estimates of maximum tsunami run-up may be based on either historical
occurrences or theoretical modeling. The Federal Insurance Administration

developed inundation maps to assist in setting rates for the National Flood
Insurance Program (flood insurance rate maps or FIRM). Inundation levels on
these maps generally consist of the higher of riverine flood, coastal storm surge, or
tsunami inundation levels. FIRM show coastal run-up elevations for 100- and 500-
year events based upon theoretical tsunami modeling for the US Pacific Coast
(Houston and Garcia 1978). Some cities, particularly in Hawaii where tsunamis
are relatively frequent, have prepared local tsunami inundation maps. These
should be available from government agencies. Eastern US cities (Atlantic and
Gulf coasts) are more likely to be affected by hurricane storm surge inundation,
whereas Pacific Coast cities are to be affected by tsunamis. Inland cities are
affected by riverine flooding, although cities adjacent to major lakes, such as
Chicago, are also susceptible to seiche. If a site lies below the inundation
elevations, then it may be vulnerable to flooding and related effects. Lower
elevations are also more vulnerable because deeper water levels carry greater
hydrodynamic forces for impact and erosion of structures and foundations.
Table 3.B-1 outlines the steps involved in assessing the potential for coastal
inundation.
3.B.4 EFFECTS OF COASTAL INUNDATION
Any of the three causes of coastal inundation (coastal subsidence, tsunami, or

seiche) results in essentially the same effect: low-lying coastal areas will be covered
with water. The greatest hazard from coastal inundation results from the inflow
and outflow of the sea during the inundation event. The strong currents from this
flow may erode foundations of structures and sweep away smaller structures or
equipment. Debris carried by the water will act as battering rams to pound other
more solidly anchored structures. People unsafely located in the low-lying areas
may be swept out to sea and drowned or knocked unconscious by the debris and
drowned. Normal wave action will also be superimposed on the tsunami, adding
dynamic forces capable of pounding and destroying even strong structures.
Petrochemical facilities would be especially vulnerable to rupture of storage tanks
from debris impact and foundation erosion, possibly resulting in explosion and
fire. The tsunami at Seward, Alaska, following the 1964 earthquake led to such
destruction of the port facilities. The succession of waves spread the fire from
ruptured oil tanks across and throughout the port area.
Tsunami combined with seiche could result in an oscillating water level
causing several cycles of coastal inundation. Periods of inundation could last for
several hours. With tectonic subsidence of the coast (or from a large submarine
landslide), the inundation may be permanent. Major subsidence, ground

SEISMIC HAZARDS 59
Table 3.B-1. Evaluation Process for Coastal Inundation Hazards.
Outline for Analysis of Coastal Inundation Potential

(1) Evaluate tsunami and earthquake history

(2) Inundation maps prepared by government agencies (e.g., FIRM, Houston
and Garcia 1978)
(3) Seismotectonic study, character of faulting, location of faulting with
respect to the site
(4) Site topography and adjacent coastal bathymetry, harbor or embayment
configuration (resonances)
(5) Hydrodynamic analyses, long-period wave response, tidal parameters
Tsunami/Seiche Inundation Evaluation Procedure
A. Determine whether tsunami/seiche/inundation is possible (history,
seismotectonic study):
1. Literature review
2. Historical research
3. Seismotectonic analysis
4. Geomorphic/paleoseismic study
5. Geotechnical study (landslides, slumps, possible generating
mechanisms)
B. Determine maximum credible run-up elevation for region:
1. Tsunami inundation maps available?
2. FIRM
3. Historical analysis
4. Theoretical analysis
C. Determine whether site is located within low-lying coastal area (below
maximum possible run-up elevation)
D. Engineering analysis required to evaluate tsunami/seiche inundation
potential
E. Engineering analysis required to identify vulnerable facility components
F. Engineering analysis required to identify/recommend specific mitigations
for vulnerable facility components
1. Seawalls or barriers
2. Facility layout to locate vulnerable components above inundation zone
3. Geotechnical studies to identify potential liquefaction/slump/landslide
hazards that might place facility at risk (Seward, Alaska, problems)
4. Facility construction with reinforced foundations and walls below
inundation levels to resist tsunami damage
5. Response plans: tsunami warning system, evacuation to high ground,
safe shutdown, automatic fire-fighting equipment(?), life-saving
(flotation) devices
G. Government studies available/needed for region?

deformation, or severe erosion could undermine bridge piers and foundations,

disrupting lifelines such as highways, railroads, pipelines, and electrical power and
communication lines. Such damage could isolate the area from emergency
equipment, supplies, and rescue.
3.B.5 MITIGATION
3.B.5.1 Shoreline Structures

Throughout Japan, along the coast where tsunami attack is common, sea walls and
other barriers have been constructed to resist tsunami inundation. These struc-
tures are sometimes unsuccessful, as large tsunamis often overtop the barriers and
flood the protected area, generating strong currents that erode the barriers and
other structures. In some instances, large wave dissipation structures have been
carried hundreds of meters inland by major tsunamis. These structures could act
as powerful battering rams against the very structures they were designed to
protect. Even large storm surf sometimes throws large boulders and rip-rap inland
causing damage to structures. Therefore, coastal protective structures must be
carefully designed to withstand extreme events.
3.B.5.2 Tsunami Warning

Tsunamis travel across the oceans at high speeds of more than 60 mi/h (100 km/
h), but generally require several hours before reaching distant coasts. Therefore,
tsunami warnings are possible and have been established through the Pacific
tsunami warning network. Citizens must take these warnings seriously: Failure to
heed the warnings by going to the beach to watch the incoming waves may result
in needless loss of life. In southern California, the great hazard that tsunamis pose
is seriously underappreciated. This perhaps results from the absence of significant
tsunamis striking the Southern California coast in the recent past (one or two
generations). Southern California beaches, such as Santa Monica, are most
vulnerable to distant tsunamis arriving from the south or north. Such tsunamis
would be generated by large or great earthquakes along the South American or
Alaska and Aleutian coasts. Other coastal areas would be vulnerable to tsunamis
from sources located along direct paths across the ocean. For large, approximately
linear tsunami sources such as oceanic trench areas, directional effects tend to
focus the tsunami energy along an axis perpendicular to the linear source region.
Sites in Hawaii are vulnerable to attack from tsunamis in many directions because
of its central location in the Pacific Ocean.
A minimum of about 20 to 30 min is required to issue a tsunami warning
through the Pacific tsunami warning network. Locally generated tsunamis would
arrive too quickly, in less than 10 to 15 min, for adequate warning by officials.
Although the technology exists to develop more rapid, regional tsunami warning
networks, the relatively rare occurrence of damaging local tsunamis and economic

SEISMIC HAZARDS 61
constraints have precluded installation of such a system in most areas (such as the
Southern California area). In the absence of official warning systems, persons in
low-lying coastal areas vulnerable to tsunami inundation should seek shelter on
higher ground immediately upon experiencing a strong earthquake in their area.

Such immediate response has been conditioned into the minds of Japanese people
where tsunami occurrences are relatively frequent. If no tsunamis were generated
in the first hour following the event, and no tsunami warnings were issued based
upon other earthquake data, returning to the low-lying area should be safe. If,
however, a tsunami is generated, the first wave may arrive and withdraw causing
little or no damage, only to be followed by larger, possibly destructive waves. For
areas that tend to have resonant oscillations, significant wave activity may persist
for several hours, or even most of a day. People in these areas should wait for an
“All Clear” signal from official agencies before returning.
3.B.5.3 Land-Use Planning and Inundation Map

Theoretical and laboratory models have been used to estimate likely inundation
levels for 100- and 500-year return periods along US coasts affected by tsunami
(FIRM). The 100- and 500-year time intervals represent an estimated average
recurrence interval; however, so-called 100-year events may occur in shorter
time spans. For example, both the 1960 Chile and 1964 Alaska earthquakes
generated tsunami waves equivalent to a 100-year event along the Southern
California coast. The run-up height estimated in these maps refers to the land
elevation (or contour) that the incoming tsunami will reach. This run-up height
includes the effects of the normal tidal range, but excludes the possible
simultaneous occurrence of the tsunami with high surf and storm wave setup.
Amplification of run-up by coastal topography may also occur, especially in
narrow, V-shaped canyons.
Sites located in vulnerable areas should prepare for the possibility of tsunami
inundation, and important structures should be designed to resist possible
inundation effects. Facilities should be evaluated to identify vulnerable compo-
nents, and possible mitigation strategies identified. Emergency or disaster re-
sponse plans should be prepared considering the vulnerability of the facility and
the possible effects of tsunami attack. Also, an attempt should be made to
anticipate otherwise unexpected consequences that may arise from the complex
interaction between direct earthquake effects and subsequent tsunami inundation.
3.B.5.4 Public Awareness

McCulloch (1985) lists the following individual actions that can help to save lives
in the event of a tsunami:
(a) If you are on low ground near the coast and a large earthquake occurs in
your area, move to high ground. There may be no time to either issue or
receive an official warning.
(b) If you recognize a drawdown of the sea, move to high ground.

(c) The first tsunami wave may not be the highest. In major tsunamis, later
waves commonly have the highest run-up heights.
(d) Periods between successive major waves may be similar. Thus, you may
have time between waves to move to higher ground or to assist in rescue

efforts.
(e) Do not assume that because the incoming tsunami wave is not breaking it
will not be destructive. The forces contained in this high-velocity, often
debris-laden torrent are extremely destructive during run-up and run-off.
(f) Do not assume that areas behind beaches generally shielded from storm
waves will be immune from high run-up. Tsunami run-up heights have
historically been higher in some such areas along the California coast.
(g) Do not go to the beach to watch a tsunami coming in. Not only might you
hamper rescue efforts, but you may also have taken your last sightseeing
trip.
References
Houston, J. R., and A. W. Garcia 1978. Type 16 flood insurance study: Tsunami prediction
for the West Coast of the Continental United States. Technical Rep. No. H-78. Vicksburg,
MS: US Army Engineer Waterways Experiment Station.
McCulloch, D.S. 1985. Evaluating tsunami potential, in evaluating earthquake hazards
in the Los Angeles region: An earth science perspective: USGS professional paper 1360,
375–413. Reston, VA: USGS.

03 - Seismic Hazards 2020

Uploaded by

Copyright:

Available Formats

03 - Seismic Hazards 2020

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

03 - Seismic Hazards 2020

Uploaded by

Copyright:

Available Formats

CHAPTER 3

Seismic Evaluation and Design of Petrochemical and Other Industrial Facilities

3.2 EARTHQUAKE BASICS

3.2.1 Earthquake Mechanism

3.2.2 Earthquake Magnitude and Intensity

Strike-Slip Faulting Reverse Faulting Normal Faulting

Figure 3-1. Examples of surface fault rupture.

Seismic Evaluation and Design of Petrochemical and Other Industrial Facilities

most commonly used magnitude scale by seismologists. Appendix 3.A presents

Earthquake intensity is a measure of the effects of an earthquake at a

3.3 GROUND SHAKING

3.3.1 Measures of Ground Shaking for Use in Design and

Seismic Evaluation and Design of Petrochemical and Other Industrial Facilities

Output time history response of an SDOF of period T1

Sa1 = Maximum response

of an SDOF of period T1 and

damping (typically 5%) plotted as a function of their period of vibration

3.3.2 Factors Affecting Ground Shaking

Seismic Evaluation and Design of Petrochemical and Other Industrial Facilities

The common approach for predicting ground motion from an earthquake at a

relationships. A ground motion prediction equation (GMPE) is a mathematical

Seismic Evaluation and Design of Petrochemical and Other Industrial Facilities

Vs30 have higher spectral accelerations at longer periods of vibration (greater

3.4 DESIGN GROUND MOTIONS

Assessment of ground motions for design includes either a computation of a

Seismic Evaluation and Design of Petrochemical and Other Industrial Facilities

In ASCE 7, MCE is deﬁned as ground motions having a 2% probability of

California. In these situations, the code allows capping the probabilistically

3.4.1 Design Response Spectrum—Standard Code Approach

Seismic Evaluation and Design of Petrochemical and Other Industrial Facilities

3.4.2 Site-Speciﬁc Design Response Spectrum

Seismic Evaluation and Design of Petrochemical and Other Industrial Facilities

Spectral Acceleration (Sa)

Figure 3-3a. Idealized design response spectrum (spectral acceleration versus

Seismic Evaluation and Design of Petrochemical and Other Industrial Facilities

Seismic Evaluation and Design of Petrochemical and Other Industrial Facilities

3.4.2.1 Probabilistic Seismic Hazard Analysis

Seismic Evaluation and Design of Petrochemical and Other Industrial Facilities

Seismic Sources Characterization

Source 1 Area Sources

Magnitude (M) Source-to-Site

Select Ground Motion Prediction Compute Hazard Curve

Distance (d) Spectral Acceleration

1. Select appropriate attenuation equation (function of source mechanism, subsurface conditions)

Hazard Deaggregation Develop Response Spectrum

Figure 3-4. Probabilistic seismic hazard analysis.

Seismic Evaluation and Design of Petrochemical and Other Industrial Facilities

3.4.2.2 Deterministic Seismic Hazard Analysis

3.4.2.3 Site Response Analysis

Seismic Evaluation and Design of Petrochemical and Other Industrial Facilities

Seismic Sources Characterization

Particular source of interest

1. Define earthquake sources and seismicity

Select Ground Motion Prediction (Attenuation) Equations

Unique magnitude-distance pair

Develop Response Spectrum

Figure 3-5. Deterministic seismic hazard analysis.

Seismic Evaluation and Design of Petrochemical and Other Industrial Facilities

(Regents of the University of California 2000) are examples of two-dimensional