Professional Documents
Culture Documents
Engineering Models of Strong Ground Motion: 5.2 The Attenuation Relation 5.3 Model Parameters
Engineering Models of Strong Ground Motion: 5.2 The Attenuation Relation 5.3 Model Parameters
5
Engineering Models of
Strong Ground Motion
5.1 Introduction
5.2 The Attenuation Relation
5.3 Model Parameters
Ground Motion Parameters · Earthquake Magnitude
· Source-to-Site Distance · Faulting Mechanism · Local Site
Conditions · Stress Drop · Source Directivity and Radiation
Pattern · Hanging-Wall and Footwall Effects · Tectonic
Environment · Instrument Location
5.4 Statistical Methods
Regression Analysis · Standard Deviation · Predicted Value
5.5 Theoretical Methods
Source Spectrum · Stress Drop · Crustal Attenuation
· Site Response · Peak Ground-Motion Parameters
5.6 Engineering Models
Western North America · Eastern North America · Europe
· Japan · Worldwide Subduction Environments · Worldwide
Extensional Environments · Hanging-Wall and Footwall Effects
· Rupture Directivity Effects
5.7 Engineering Evaluation
Defining Terms
Kenneth W. Campbell References
ABS Consulting and EQECAT, Inc.
Further Reading
Portland, Oregon Appendix–Notation
5.1 Introduction
There are two basic methods used to estimate strong ground motion in engineering practice. In the
first method, known as deterministic seismic hazard analysis or DSHA, ground motion is estimated from
a given set of seismological parameters, such as earthquake magnitude and the distance from the earth-
quake rupture zone to the site of interest. In the second method, referred to as probabilistic seismic
hazard analysis or PSHA, ground motion is estimated statistically using all possible earthquake locations
and magnitudes together with their expected probabilities of occurring. Both of these methods have one
thing in common: they need a means of estimating strong ground motion from the specified seismological
parameters. This estimation is usually done using a ground motion relation, or what is commonly referred
to in engineering as an attenuation relation.
Only engineering models that directly predict ground-motion amplitude or that predict the modula-
tion of this amplitude from such effects as fault geometry and source directivity are discussed in this
chapter. Duration is an important aspect of strong ground motion, especially for the inelastic response
of structures, but much less attention has been paid to predicting duration and, therefore, no consensus
engineering models are available. Generally speaking, the inelastic behavior of structures is included in
structural design through the use of time histories and structural ductility factors, which are the topics
of other chapters in this book.
Although it is not the intent of this chapter to provide a history of engineering models, it is important
to mention some of the pioneers whose early and continued efforts led to their widespread acceptance
in engineering and engineering seismology, particularly for use in the then-fledgling field of PSHA. Some
of these pioneers (listed alphabetically) include Neville Donovan, Luis Esteva, George Housner, Ed Idriss,
Bill Joyner, Kiyoshi Kanai, Robin McGuire, H. Bolton Seed, and Mihailo Trifunac. Many others, including
the author, have followed in the footsteps of these pioneers, and this body of work forms the basis for
the contemporary engineering models discussed in this chapter. A more complete review can be found
in the references provided in the section entitled Further Reading.
Y = c 1e c 2 M R − c 3 e − c 4R e c 5 F e c 6Se (5.1)
ln Y = c 1 + c 2 M − c 3 ln R − c 4 R + c 5 F + c 6 S + ε (5.2)
r + c 7 exp (c 8 M ) or
R= 2 (5.3)
[
r + c 7 + exp (c 8 M ) ]
2
In the above equations, Y is the strong-motion parameter of interest, M is magnitude, F is the faulting
mechanism of the earthquake, S is a description of the local site conditions beneath the site, ε is a
random error term with a mean of zero and a standard deviation of σln Y (the standard error of estimate
of ln Y), and r is a measure of the shortest distance from the site to the source of the earthquake. In the
more complicated forms of Equations 5.1 to 5.3, the coefficients c3, c6, and c7 are defined in terms of M
and R. Many of these coefficients also have been found to be dependent on the tectonic environment
of the regions in which the earthquakes occurred. Figure 5.1 shows an attenuation relation for peak
ground acceleration (PGA) and the database on which it was based. A selected set of attenuation relations
that are commonly used for engineering evaluation and design throughout the world are presented later.
Many of the mathematical expressions in Equations 5.1 to 5.3 have their roots in earthquake seismology
[Lay and Wallace, 1995]. For example, the expressions Y ∝ e c2M and ln Y ∝ c2M are consistent with the
original definition of earthquake magnitude [Richter, 1935]. The expressions Y ∝ R –c3 and ln Y ∝ –c3
ln R are consistent with the geometric attenuation of the seismic wave front as it propagates away from
the earthquake source. Some attenuation relations assume c3 = 1, which is the theoretical value for
spherical spreading of the wave front from a point source in a homogeneous whole space. If uncon-
strained, c3 typically will be greater than 1. Sometimes c3 is varied as a function of distance to accom-
modate differences in the geometric attenuation of different wave types, such as direct waves or surface
waves, and to account for the critical reflection off the base of the crust or other strong crustal reflectors.
The expressions Y ∝ e –c4R and ln Y ∝ –c4R are consistent with the anelastic attenuation that results from
material damping and scattering as the seismic waves propagate through the crust. The relationship
Uncorrected PGA
Corrected PGA
1
0.1
M 5.5
8.0
Corrected
Additional Uncorrected
7.5
Moment Magnitude
7.0
6.5
6.0
5.5
5.0
4.5
0 10 20 30 40 50 60
Distance to Seismogenic Rupture (km)
FIGURE 5.1 Example PGA attenuation relation (top) and its associated database (bottom). Uncorrected recordings
are analog or digital acceleration time histories that have not been processed and, therefore, can provide only estimates
of PGA. Corrected recordings are acceleration times histories that have been processed to derive velocity and
displacement time histories, response spectra, and Fourier amplitude spectra. (From Campbell, K.W. and Bozorgnia,
Y. 1999. “Vertical Ground Motion: Characteristics, Relationship with Horizontal Component, and Building-Code
Implications,” in Proc. SMIP99 Seminar on Utilization of Strong-Motion Data, M. Huang, Ed., Sept. 15, San Francisco,
pp. 23–49. California Strong Motion Instrumentation Program, Sacramento. With permission.)
between Y and the remaining parameters have been established over the years from both empirical and
theoretical ground-motion modeling.
The mathematical relationships defined in Equation 5.3 are used to incorporate the widely held belief
that ground motion at short period or high frequency should become less dependent on magnitude (or
saturate) close to the causative fault. Schnabel and Seed [1973] first modeled this behavior using simple
geometrical considerations. Since then, similar results have been obtained using empirical models [Camp-
bell, 1981] and theoretical finite-source models [Anderson, 2000]. Now, most modelers consider this an
accepted behavior of near-fault ground motion.
The remainder of this chapter describes in some detail the strong-motion and seismological parameters
that appear in Equations 5.1 to 5.3. The discussion specifically addresses issues regarding the use of
attenuation relations to predict strong ground motion for engineering evaluation and design. Particular
emphasis is placed on the prediction of peak time-domain and peak frequency-domain parameters,
because these parameters are those most commonly used by engineers. This is followed by a selected
compilation of recently published attenuation relations.
2
2π 2π
PSA = PSV = SD (5.4)
Tn Tn
9
MS
W
M
MJMA
~
M
8
mB
7 ML
mb
MAGNITUDE
4
S
M
L
M
2
2 3 4 5 6 7 8 9 10
MOMENT MAGNITUDE
FIGURE 5.2 Comparison of magnitude scales. (From Heaton, T.H., Tajima, F., and Mori, A.W. 1986. “Estimating
Ground Motions Using Recorded Accelerograms,” Surv. Geophys., 8, 25–83. With permission.)
regional estimates of magnitude due to differences in the way it is locally defined and calculated. The
definition of all peak time-domain magnitude scales can be given by the equation [Lay and Wallace, 1995]:
( )
M = log ( A T ) + f R, hhypo + C s + Cr (5.5)
where A is the amplitude of the seismic phase on which the magnitude scale is based measured on
a seismometer, T is the period of the signal, f(R, hhypo ) is a correction for the distance from the
earthquake source to the seismometer and the hypocentral depth hhypo, Cs is a correction for the siting
of the seismometer, and Cr is a correction for the earthquake source region.
There is an increasing tendency to adopt MW as the worldwide standard for quantifying magnitude
because of its strong physical and seismological basis [Bolt, 1993]. By definition, MW is related to seismic
moment M0 , a measure of the seismic energy radiated by an earthquake, by the formula [Hanks and
Kanamori, 1979; Kanamori, 1978]:
2
MW = log M 0 − 10.7 (5.6)
3
where
M 0 = µA f D = 2 µES ∆σ
where
µ = shear modulus of the crust in the source region
Af = rupture area
D = average displacement on the rupture plane
∆σ = stress drop
Es = radiated seismic energy
The definition based on Af D allows M0 to be derived from geological faulting parameters that can be easily
observed in the field for large surface-rupturing earthquakes. The definition based on Es /∆σ allows M0 to
be derived from seismological measurements.
rhypo = repi
2
+ hhypo
2
(5.7)
where hhypo is the depth of the hypocenter, measured from the Earth’s surface. Generally speaking,
repi and rhypo are poor measures of distance for earthquakes with large rupture areas. They are primarily
used for characterizing distances for small earthquakes that can be reasonably represented by a point
source. Experience has shown that attenuation relations that use point-source measures should not
be used to estimate ground motions close to large earthquakes unless there is absolutely no other
alternative available.
There are three finite-source distance measures that are commonly used in practice: rjb or the closest
horizontal distance to the vertical projection of the rupture plane, introduced by Joyner and Boore [1981];
rrup or the closest distance to the rupture plane, introduced by Schnabel and Seed [1973]; and rseis or the
closest distance to the seismogenic part of the rupture plane, introduced by Campbell [1987, 2000b].
The measure rseis assumes that fault rupture within the near-surface sediments or shallow fault gouge is
nonseismogenic, which was later proposed and demonstrated seismologically by Marone and Scholz
[1988]. These distance measures are compared in Figure 5.3.
Although rjb is reasonably easy to estimate for a future (design) earthquake, rrup and rseis are not as
easily determined, particularly when the earthquake is not expected to rupture the entire seismogenic
width of the crust. In such cases, the average depth to the top of the inferred rupture plane, drup , or to
the seismogenic part of this rupture plane, dseis, can be calculated from the equation [Campbell, 2000c]:
1
[
H + H bot − W sin (δ)
di = 2 top ] for di ≤ H i
(5.8)
H i otherwise
where the subscript i = rup or seis is the distance measure of interest, Hbot is the depth to the bottom
of the seismogenic part of the crust, Htop is the depth to the top of the fault, Hseis is the depth to the
top of the seismogenic part of the fault, δ is the angle between the Earth’s surface and a line extending
Vertical Faults
rjb
rrup
rseis
Seismogenic
Depth
rhypo
Hypocenter
Dipping Faults
rjb = 0 rjb
rrup
rseis
Seismogenic
rhypo Depth rseis & rrup rhypo
Hypocenter Hypocenter
FIGURE 5.3 Comparison of distance measures. (From Abrahamson, N.A. and Shedlock, K.M. 1997. “Overview,”
Seismol. Res. Lett., 68, 9–23. With permission.)
down-dip perpendicular to the strike of the fault (dip angle); and W is the down-dip width of the
fault rupture.
The rupture plane width can be calculated from the expression [Wells and Coppersmith, 1994]:
TABLE 5.1 Representative Values for the Average Depth to the Top of Fault Rupture
drup (km) dseis (km)
MW W (km) δ = 30° δ = 45° δ = 90° δ = 30° δ = 45° δ = 90°
is a continuous variable representing the angle between the direction of slip on the fault plane and
the strike or the orientation of the fault on the Earth’s surface. Rake angle has not been used directly
in an attenuation relation to define faulting mechanism. Instead, the faulting mechanism has been
classified in terms of two or more categories. These categories include strike slip, reverse slip, and
normal slip (Chapter 4). The values of rake angle corresponding to these faulting mechanisms are 0°
for left-lateral strike-slip faulting, 180° for right-lateral strike-slip faulting, 90° for reverse faulting,
and 270° for normal faulting [Lay and Wallace, 1995]. Thrust faulting is a special case of reverse
faulting in which the dip angle of the rupture plane is less than 45°. A combination of strike-slip with
either reverse-slip or normal-slip is known as oblique faulting and will have a rake angle that falls
between those given here.
Campbell [1981] empirically demonstrated that reverse and thrust faulting causes higher ground
motion than strike-slip or normal faulting. Many subsequent studies have proven this effect to be
universal. It has been common practice in the past to put strike-slip and normal-faulting events into a
single category. However, a recent study by Spudich et al. [1999] suggests that normal-faulting events,
or strike-slip events in an extensional stress regime, may have lower ground motions than other types
of shallow crustal earthquakes.
There has been a great deal of interest in blind thrust faults after observing unusually large ground
motions during the 1987 Whittier-Narrows, CA earthquake, the 1988 Saguenay, Canada earthquake, and
the 1994 Northridge, CA earthquake. Whether similarly high ground motions can be expected from all
future blind-thrust earthquakes is speculative at present. However, it cannot be ruled out, considering
the current limited observational database. The higher ground motions observed during the previously
mentioned earthquakes have been found to correspond to higher-than-average stress drop. More theo-
retical and empirical studies will be needed before there is a clear understanding why these earthquakes
produced such high stress drops and how such events can be predicted in the future.
Source: Adapted from Wills, C.J. et al. 2000. “A Site-Conditions Map for Califor-
nia Based on Geology and Shear-Wave Velocity,” Bull. Seismol. Soc. Am., 90,
S187–S208. With permission.
ground motion propagates from below see (Chapter 6). A general description of the different site classes
used in recently published attenuation relations is presented later and in the Appendix.
Traditionally, local site conditions have been classified simply as soil or rock. Many attenuation
relations continue to use this simple classification. Campbell [1981] proposed that site conditions should
be subdivided into shallow soil, soft soil, Holocene or firm soil, and Pleistocene or very firm soil, and
that rock should be further subdivided into soft or primarily sedimentary rock and hard or primarily
crystalline rock. Although this more refined geological classification has not been utilized in most
attenuation relations, Campbell and Bozorgnia [in press] clearly demonstrated the importance of this
classification scheme in the prediction of near-source ground motion. Park and Elrick [1998] and Wills
and Silva [1998] have also shown that a more refined geological classification is warranted based on
measurements of shear-wave velocity in various geologic units in California.
There are typically two methods for classifying a site in terms of shear-wave velocity, VS . The first is
the average value of VS in the top 30 m (100 ft) of a site profile, referred to as 30-m velocity. The second
is the average value of VS over a depth equal to a quarter-wavelength of the period or frequency of
interest, referred to by Boore and Joyner [1991] as effective velocity. Modern U.S. building codes [BSSC,
1998; ICBO, 1997; ICC, 2000] have all adopted the 30-m velocity, designated VS30 in this chapter, as the
primary basis for classifying a site for purposes of incorporating local site conditions in the estimation
of design ground motion. These building codes define five site classes in terms of a range of 30-m
velocities, designated A through E. Wills et al. [2000] extended this classification in California to include
intermediate values. Typical values (usually the midpoint of the ranges) are given in Table 5.2. For the
site classes defined by inequalities, typical values for Soft Soil are from Borcherdt [1994] and those for
Hard Rock, considered to be an older sedimentary rock in eastern North America, are from Savy and
coworkers [1987].
The value of 30-m velocity is determined from the formula:
n n
VS 30 = ∑ ∑d
i =1
di
i =1
i VSi (5.10)
where di is the thickness and VSi is the shear-wave velocity of soil layer i. Progressively deeper soil layers
are incorporated until the summation in the numerator equals 30 m (100 ft). Boore and coworkers [1993]
were the first to use site categories based on VS30 in the development of an attenuation relation. In 1994,
these same authors were the first to use VS30 directly as a parameter in an attenuation relation [Boore et
al., 1994]. Because of its adoption in the building codes, 30-m velocity has become the preferred site
parameter in engineering analysis. According to Boore and Joyner [1997], VS30 = 310 m/sec and
VS30 = 620 m/sec are reasonable estimates of 30-m velocity for generic soil and generic rock sites in western
North America.
Joyner and coworkers [1981] proposed a VS-based site parameter that is related to the nonresonant
amplification produced as a result of the energy conservation of seismic waves propagating vertically
upward through a site profile of gradually changing velocity. This parameter, later referred to as effective
velocity [Boore and Joyner, 1991], is defined as the average velocity from the surface to a depth corre-
sponding to a quarter-wavelength of the period or frequency of interest. Effective velocity can be calcu-
lated from Equation 5.10 by summing to a depth corresponding to a quarter-wavelength rather than
30 m. This depth is given by the expression [Boore, in press]:
D1/ 4 ( f ) = ∑d
i =1
i (5.11)
where T = 1/f is the period of interest. Progressively deeper soil layers are used in the above sum-
mation until the equality
∑d
i =1
i VSi = T 4
is met.
Effective velocity should be a better site parameter than 30-m velocity because it takes the period of
the wave into account. Joyner and coworkers [Joyner and Fumal, 1984; Joyner and Boore, 1988] are the
only investigators to include effective velocity as a parameter in an empirical attenuation relation. Effective
velocity has found widespread use in the calculation of site amplification using the stochastic method
discussed later. The two attenuation relations for the eastern United States presented later were developed
using the stochastic method.
Sediment or basin depth is the depth to the basement-rock horizon beneath the site. Basement rock
is a geological term that is used to describe the more resistant, generally crystalline rock that lies beneath
layers or irregular deposits of younger, relatively deformed sedimentary rock. It was introduced as a site
parameter by Trifunac and Lee [1979] and later used by Campbell [1987, 2000b] to quantify the response
of long-period ground motion. Both of these investigators have continued to use this parameter in their
subsequent studies but it has not generally been used by others. Recently its importance has been
recognized by several seismologists. For example, based on empirical and theoretical considerations,
Joyner [2000] found that sediment depth appeared to be a reasonable proxy for modeling the effects of
traveling surface waves generated at the edge of a sedimentary basin. Lee and Anderson [2000] and Field
and the SCEC Phase III Working Group [2000] found that sediment depth could be used to approximately
model the three-dimensional response of the Los Angeles basin. Rodriguez-Marek and coworkers [2001]
found that depth to bedrock, defined as VS ≥ 760 m/sec, was an important parameter in estimating
seismic site response from the 1989 Loma Prieta and 1994 Northridge, CA earthquakes.
FIGURE 5.4 Radiation pattern showing the variability of compressional and horizontal shear-wave amplitude for
a fault rupture propagating from left to right. The diagrams on the left are for a rupture propagation velocity of 0.5
times the shear-wave velocity and those on the right are for a rupture propagation velocity of 0.9 times the shear-
wave velocity. (From Lay, T. and Wallace, T.C. 1995. Modern Global Seismology, Academic Press, San Diego. With
permission.)
relatively low ground motions during the Chi-Chi earthquake is particularly significant, because it was
a large thrust earthquake, which had been expected from previous empirical and theoretical studies to
have relatively large ground motions. The relatively low stress drops implied for the Taiwan and Turkey
earthquakes may have been caused by the large total slip on the causative faults [Anderson, in press] or
because they ruptured the Earth’s surface [Somerville, 2000]. More study will be needed to better
understand the phenomena that may have contributed to these low ground motions. If these earthquakes
are found to be typical of similar large earthquakes worldwide, then the implication is that the current
attenuation relations overpredict short-period ground motions from large earthquakes, an attribute
suggested from observations of precarious rocks near great earthquakes on the San Andreas fault [Brune,
1999].
FIGURE 5.5 Radiation pattern for a vertical strike-slip fault showing its effect on the fault-normal and fault-parallel
components of near-fault ground displacement. (From Somerville, P.G., Smith, N.F., Graves, R.W., and Abrahamson,
N.A. 1997. “Modification of Empirical Strong Ground Motion Attenuation Relations to Include the Amplitude and
Duration Effects of Rupture Directivity,” Seismol. Res. Lett., 68, 199–222. With permission.)
Rupture directivity is a well-known seismological principle [Lay and Wallace, 1995]. It has been
observed or proposed as a factor in controlling the azimuthal dependence of strong ground motion
during the 1979 Imperial Valley earthquake [Singh, 1985], the 1980 Livermore earthquakes [Boatwright
and Boore, 1982], the 1989 Loma Prieta earthquake [Campbell, 1998], the 1992 Landers earthquake
[Campbell and Bozorgnia, 1994], and the 1999 Chi-Chi, Taiwan and Kocaeli, Turkey earthquakes [Som-
erville, 2000]. Except to a limited extent [Campbell, 1987, 2000b], source directivity has not been used
directly in the development of an attenuation relation. Somerville and coworkers [1997] and Abrahamson
[2000] developed a simple engineering model, presented later, for estimating the effects of source
directivity and radiation pattern on the prediction of the fault-normal and fault-parallel components of
PGA and PSA. Somerville and coworkers [1997] provide a list of near-source time histories that they
believe contain significant directivity and other near-source effects that can be used for engineering
evaluation and design.
Somerville [2000] suggests that the empirical models proposed by Somerville et al. [1997] and Abra-
hamson [2000] are likely to be too simplistic. He has found that the near-fault directivity effects observed
in recent earthquakes, including the 1999 Chi-Chi, Taiwan and Kocaeli, Turkey events, appear to manifest
themselves as narrow-band pulses whose period increases with increasing magnitude. This increase in
period can actually lead to lower values of PSA at mid-periods (Tn ≈ 1 sec) for MW > 7 1/4 . This latter
observation is inconsistent with the assumption made in the current engineering model that spectral
amplitudes must increase monotonically with period. The directivity pulse model needs more develop-
ment before it is ready to be used in engineering. Until then, the simple engineering model suggested
by Somerville et al. [1997] and Abrahamson [2000] will remain the state of the art.
FIGURE 5.6 Geographic distribution of active and stable continental tectonic regions worldwide. (From Johnston,
A.C. 1994. “Seismotectonic Interpretations and Conclusions from the Stable Continental Region Seismicity Database,”
in The Earthquake of Stable Continental Regions, Vol. 1, Assessment of Large Earthquake Potential, Electric Power
Research Institute, Palo Alto, CA, pp. 1–103. With permission.)
5-15
© 2003 by CRC Press LLC
0068_C05_fm Page 16 Thursday, August 1, 2002 7:59 AM
analysis is used to determine the best estimate of the coefficients c1 through c8 in Equations 5.2 and 5.3
using statistical fitting procedures such as minimum least squares or maximum likelihood. Traditionally,
there have been three methods for performing a regression analysis for the purpose of developing an
attenuation relation: (1) weighted nonlinear least-squares regression, introduced by Campbell [1981];
(2) two-step regression, introduced and later refined by Joyner and Boore [1981, 1994]; and (3) random-
effects regression, introduced by Brillinger and Preisler [1984] and later refined by Abrahamson and
Youngs [1992]. Each of these methods has its strengths and weaknesses but they all have the same intended
purpose — to mitigate the bias introduced by the uneven distribution of recordings with respect to
magnitude, distance, and other seismological parameters. The advantage of the latter two methods is
that they provide a direct estimate of the intra- and inter-earthquake components of randomness.
∑ (ln Y − ln Y )
1 2
σ lnY = (5.12)
n− p i =1
i i
where n is the number of observations, p is the number of regression coefficients, i is the index of
the recording, ln Yi is the observed value, and lnYi is the predicted value. Plots of the residuals vs.
one or more seismological parameters such as magnitude or distance are used to identify unwanted
biases in the attenuation relation by revealing trends in the residuals. Any statistically significant
trend identified visually or from a formal statistical hypothesis test [Draper and Smith, 1981] can
indicate epistemic or modeling uncertainty, such as a problem with the functional form of the
relation or the need to include another seismological parameter in the model [Campbell, 1985].
Figure 5.8 gives an example of residual plots that do not indicate a significant trend with magnitude
or distance and, therefore, indicate an unbiased model.
It is often convenient to segregate σlnY into its intra- or within-earthquake and its inter- or between-
earthquake components, traditionally designated σ and τ. Specific algorithms for determining σ and τ
are given by Abrahamson and Youngs [1992] and Joyner and Boore [1993, 1994] for the random-effects
and two-step regression procedures. If the regression analysis is performed on the geometric mean
(average of the logarithms) of the two horizontal components of a strong-motion parameter and the
variability between these components is desired, an additional component-to-component standard devi-
ation term is needed. The predicted value of strong ground motion that includes component-to-com-
ponent variability is what Joyner and Boore [1993, 1994] refer to as the random horizontal component.
Referring to these thr ee standar d deviations as σintra, σinter , and σcomp , the total standar d deviation of lnY
is given by the expression:
σ lnY = σ inter
2
+ σ intra
2
+ σ 2comp (5.13)
The standard deviation of lnY has been found to be a function of magnitude [e.g., Youngs et al., 1995]
and the amplitude of ground motion [Donovan and Bornstein, 1978; Campbell and Bozorgnia, 1994].
However, these dependencies are not always taken into account when determining the value of σlnY .
These dependencies can be significant and will always result in lower values of σlnY for larger magnitudes
and higher amplitudes. Lower estimates of standard deviation will always lead to a significant reduction
in the higher percentile estimates of ground motion computed using a DSHA and in the longer return-
period and higher percentile estimates of ground motion computed using a PSHA.
4 All Faults
All Soils
3
Normalized Residual 2
-1
-2
-3
-4
-5
4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0
Moment Magnitude
4 All Faults
All Soils
3
Normalized Residual
-1
-2
-3
-4
-5
0.0 10.0 20.0 30.0 40.0 50.0 60.0
Distance to Seismogenic Rupture (km)
FIGURE 5.8 Residuals from a regression analysis of strong-motion data plotted against magnitude (top) and
distance (bottom). (From Campbell, K.W. and Bozorgnia, Y. 1999. “Vertical Ground Motion: Characteristics, Rela-
tionship with Horizontal Component, and Building-Code Implications,” in Proc. SMIP99 Seminar on Utilization of
Strong-Motion Data, M. Huang, Ed., Sept. 15, San Francisco, pp. 23–49. California Strong Motion Instrumentation
Program, Sacramento. With permission.)
σ ln2 Y
ln Y1−α = ln Y + t v (α ) + σ ln2 Y (5.14)
n0
where tv(α) is the Student’s t-statistic for a given exceedance probability α and v = n – p degrees of
freedom (this statistic is widely available in statistics books), and σ lnY is the standard error of the
mean value of ln Y excluding random error (an indicator of epistemic uncertainty). The 100(1 − α)-
percentile estimate of a single future observation of ln Y, the most common application of Equation
5.14, is calculated by letting n0 = 1.
It is common to calculate the 100(1 − α)-percentile estimate of a single future value of lnY by setting
σ lnY = 0 and replacing the t-statistic with the standard normal variable z. These assumptions reduce
Equation 5.14 to its more simplified and common form:
where zα is the standard normal variable for an exceedance probability of α (this variable is widely
available in statistics books). Although statistically incorrect, results using Equation 5.15 are not
significantly different from those using Equation 5.14, except when the predicted value is based on
an extrapolation of the regression equation or on very few recordings. In the first case, the value of
σ lnY should not be ignored, and in the second, the z-statistic is inappropriate.
The most common application of Equation 5.15 in DSHA is to estimate either the median value of
Y (α = 0.5), by setting zα = 0, or the 84th-percentile value of Y (α = 0.16), by setting zα = 1. The 84th-
percentile value is often used as a conservative estimate of ground motion to use in the design of critical
facilities, such as nuclear power plants and major dams. Equation 5.15 is assumed to be a continuous
distribution of Y when used in PSHA, although it is usually truncated at zα = 2 or 3 (i.e., at ±2 to 3
standard deviations) to avoid unrealistic predicted values. In both DSHA and PSHA, it is typical to
include epistemic uncertainty in the predicted ground motion by using more than one attenuation
relation to estimate Y in lieu of formally defining σ lnY .
where Src (f ) describes the earthquake source, Attn (f ) describes the attenuation caused by wave
propagation through the crust, and Amp(f) describes the response of the materials beneath the site.
2 ℜ V FS
Src ( f ) = (2 π f ) 3
M0S ( f ) (5.17)
4 πρβ
where R = 0.55 is the average shear-wave radiation pattern, V = 1 2 is the partition of the radiated
energy into two orthogonal horizontal components, FS = 2 is the amplification due to the free
surface, β is the shear-wave velocity of the crust in the source region, ρ is the density of the crust in
the source region, M0 is seismic moment, and S(f) is the source displacement spectrum, known
simply as the source spectrum.
The simplest representation of the source spectrum is the one-corner, point-source model [Brune
1970, 1971]:
1
S( f ) = (5.18)
1+( f f0)
2
where f0 is corner frequency (the inflection point in the source spectrum). Joyner [1984] argued on
theoretical grounds that the source spectrum should have two corner frequencies, if the width of
faulting is limited by the finiteness of the seismogenic crust. Since then, others have developed two-
corner source models based on both empirical and theoretical considerations. All of these models are
based on the addition or multiplication of two one-corner source spectra, similar to Equation 5.18,
with each defined by a separate corner frequency, fa and fb, and combined using a relative weighting
factor. Atkinson and Boore [1998] give a comparison of the near-source acceleration source spectra
estimated from several of these one-corner and two-corner models. This comparison is shown in
Figure 5.9.
There are two attenuation relations presented later that are based on the stochastic method. One uses
a one-corner source spectrum and the other a two-corner source spectrum. The one-corner source
spectrum used in the Toro and coworkers’ [1997] attenuation relation is given by Equation 5.18, in which
the corner frequency is related to magnitude by the expression [Atkinson and Boore, 1998]:
assuming ∆σ = 120 bar, hhypo = 10 km, and β = 3.76 m/sec for the Midcontinent region and 2.58 km/
sec for the Gulf Coast region [EPRI, 1993].
According to Atkinson and Boore [1998], the two-corner source spectrum used in the Atkinson and
Boore [1995, 1997] attenuation relations is given by the expression:
1− w w
S( f ) = + (5.20)
1+( f fa ) 1+( f fb )
2 2
where
104
103 M = 7.5
Fourier Acceleration Spectrum (cm/s)
102
M = 4.5
101
AB95
100 AB98-Ca
H96
Fea96 (no site amp)
BC92
J97
10−1
10−2 10−1 10−0 101 102
freq (Hz)
FIGURE 5.9 Fourier amplitude spectra of acceleration for a reference distance of 1 km, based on several source
spectral models. (From Atkinson, G.M. and Boore, D.M. 1998. “Evaluation of Models for Earthquake Source Spectra
in Eastern North America,” Bull. Seismol. Soc. Am., 88, 917–934. With permission.)
The differences in ground motion between the one-corner and two-corner source spectra can be
significant. For example, the two-corner spectrum has been shown to produce PSA values in the mid-
period range that are substantially smaller than those predicted by the one-corner model [Atkinson and
Boore, 1995, 1997].
f 0 = 4.9 × 106 β ( ∆σ M 0 )
1/ 3
(5.24)
seismologists prefer to call the Brune stress drop the “stress parameter” [Atkinson and Beresnev, 1997].
The term stress drop is retained here to be consistent with Brune’s original terminology.
Attn ( f , R) = G ( R) D ( f ) (5.25)
where G(R) describes the geometric attenuation of ground motion and D (f ) describes the diminution
of ground motion with distance from the source, or crustal damping. Geometric attenuation is
modeled using the distance parameter R−n, where R is either rhypo or rrup, and n varies with distance
according to the dominant wave type (e.g., direct or surface waves) and whether arrivals of critical
reflections off the Moho (the top of the Earth’s mantle) or other strong crustal reflectors are
included. Spherical spreading from a point source corresponds to n = 1.
In the Toro and coworkers’ [1997] attenuation relation presented later, geometric attenuation is given
by the general expression:
1/ 2
n
G ( R) =
∑
i =1
G ( R)
i
2
(5.26)
where Gi is the geometric attenuation of an individual ground-motion ray in a layered crust deter-
mined from geometrical optics and n is the total number of arrivals within the ergodic window [Ou
and Herrmann, 1990]. In the Atkinson and Boore [1995, 1997] attenuation relation presented later,
geometric attenuation is given by the three-part expression:
1 R for R < 70 km
G ( R) = 1 70 for 70 ≤ R < 130 km (5.27)
(1 70) 130 R for R ≥ 130 km
−π f R
D ( f ) = exp (5.28)
Q ( f )β
where Q (f ), the quality factor, is a measure of anelastic attenuation and scattering within the crust,
which is usually defined by the power law:
Q ( f ) = Q0 f η
(5.29)
where Q0 and η vary according to the tectonic environment [e.g., Singh and Herrmann, 1983].
ρβ
Amp ( f ) = exp (− πκ 0 f ) (5.30)
ρs ( f )β s ( f )
where β and ρ are the shear-wave velocity and density of the crust beneath the site. The effective
velocity βs (f ), effective density ρs (f ), and site damping factor κ0 are calculated from the expressions
[Boore, in press]:
n1 n1
∑ ∑v
di
βs ( f ) = di (5.31)
i =1 i =1 si
n1 n1
ρs ( f ) = ∑
i =1
ρi di ∑di =1
i (5.32)
n2
∑v Q
di
κ0 = (5.33)
i =1 si si
where I(f, fn ) is given by [Silva and Lee, 1987; Liu and Pezeshk, 1999]:
f n2
I ( f , fn) = 1/ 2
(5.35)
(
f2−f2
) + (2ξ ff )
2 2
n c n
The expected value of PGA or PSA is calculated from A(f) or A(f, fn) using random vibration theory
(RVT). In RVT, the peak of a random function is calculated from its root-mean-square (RMS) value
according to the relationship [Cartwright and Longuet-Higgins, 1956]:
Y = YRMS f ( N ) (5.36)
where the peak factor f (N) ≈ 2 ln N + γ Euler 2 ln N , when N (the number of cycles in the time
domain) is large, and γEule r = 0.55726 (Euler’s constant). From Parseval’s theorem, the RMS acceler-
ation is defined by the integral:
1/ 2
∞
∫ Y(f )
1
= df
2
YRMS (5.37)
DRMS
−∞
where Y (f ) = A (f ) or A(f, fn), depending on whether Y is PGA or PSA, and DRMS is the equivalent
RMS duration. One of the critical elements of this relation is the appropriate selection of DRMS. This
is particularly critical for SDOF oscillators with long periods where, for small earthquakes, the natural
period can be longer than the duration of the ground motion. This issue is addressed by Boore and
Joyner [1984] and Liu and Pezeshk [1999]. The expression proposed by Liu and Pezeshk [1999] is
given by:
γ2
DRMS = DS + D0 2 (5.38)
γ +k
where
γ = DS Tn (5.39)
D0 = Tn 2πξ c (5.40)
m12
k = 2π 1 − (5.41)
m0m 2
∞
2π
mi =
DS ∫ f G ( f )df
0
i
(5.42)
and DS is the duration of the source often assumed to be 1/f0 + 0.05R [Herrmann, 1985], mi is the
ith spectral moment of the power spectral density function of the random process, and G(f) is the
one-sided power spectral density function.
Boore [1996] distributes a set of computer programs that can be used to evaluate the stochastic method
using a point-source representation of the earthquake source. These programs can be used to estimate
peak parameters using RVT or to simulate ground-motion time histories. Beresnev and Atkinson [1998]
distribute a computer program that can be used to evaluate the stochastic method assuming the earth-
quake is a finite rather than a point source.
Western North America Shallow active crust Abrahamson and Silva [1997]
Boore et al. [1997]
Campbell and Bozorgnia [in press]
Sadigh et al. [1993, 1997]
Eastern North America Shallow stable crust Atkinson and Boore [1995, 1997]
Toro et al. [1997]
Campbell [in press]
Europe Shallow active crust Ambraseys et al. [1996]
Shallow stable crust Dahle et al. [1990]
Japan All types undivided Molas and Yamazaki [1995, 1996]
Worldwide Shallow extended crust Spudich et al. [1999]
Subduction interface Youngs et al. [1997]
Subduction intraslab Youngs et al. [1997]
Subduction undivided Crouse [1991a, 1991b]
in this case fraction of g for acceleration and cm/sec for velocity. Pseudovelocity was converted to
pseudoacceleration if the original relation was given in terms of PSV. Logarithms are consistently defined
in terms of the natural logarithm in the attenuation relations. All other equations were left in terms of
the common logarithm if that is how they were defined originally.
Some of the model’s functional forms were modified in order to simplify or generalize them so that
a single table of regression coefficients could be used. Although this means that the attenuation relations
depart from their original mathematical forms, the modifications provide a consistent framework for
understanding and applying these relations in an informed and unambiguous manner. The strong-
motion parameters are defined as either the largest horizontal component or the geometric mean
(hereafter referred to simply as the average) of the two horizontal components, with no attempt to convert
between the two. If such a conversion is necessary, the largest horizontal component can be estimated
approximately from the average horizontal component by multiplying by 1.15 [Campbell, 1981; Ansary
et al., 1995].
All engineering models have limitations that result from the availability of recordings, the criteria used
to select the recordings, the theoretical assumptions used to develop the models, and the seismological
parameters used to define the source, path, and site effects. It is dangerous to assume that an engineering
model can be extrapolated beyond the data, the theoretical assumptions, or the geographic region of
applicability, and still provide a reliable estimate of ground motion. In fact, some of these models come
with specific caveats regarding their use. When such caveats are given, they have been noted.
The notation used to define the parameters in the engineering models is defined in the Appendix so
that it need not be repeated after each model. Two specific parameters, site class and faulting mechanism,
generically designated S and F, pose a particular challenge, because each model defines these parameters
somewhat differently. Different site classes used in the attenuation relations are denoted by different
subscripts. An approximate correspondence among values of VS30, the related site classes, and the site
parameters used in the attenuation relations given in Table 5.3 is listed in Table 5.4. The faulting
mechanism parameter is generally set to 0, 0.5, or 1, depending on the particular faulting mechanism
category but each model defines these categories differently in terms of the rake angle λ. Instead of
defining different faulting mechanism parameters for each model, the values of F and the corresponding
range of λ for the attenuation relations given in Table 5.3 are listed in Table 5.5.
Note: Approximate values of 30-m velocity for each site class in m/sec are listed under the entries for Boore et al. [1997]. Blank entries refer to parameters that are not used for that attenuation
relation.
5-25
© 2003 by CRC Press LLC
0068_C05_fm Page 26 Thursday, August 1, 2002 7:59 AM
TABLE 5.5 Faulting Mechanism Categories and Related Rake Angles for Selected Attenuation
Relations
Attenuation Relation Category F Rake Angle (λ)
Note: Unless otherwise indicated, an unknown or random faulting mechanism is given by F = 0.5,
FRV = 0.25, and FTH = 0.25.
assumed that the WNA attenuation relations, some of which use a limited number of recordings outside
of North America, can be used to estimate ground motions in similar tectonic environments worldwide.
This has been confirmed to a limited extent by comparing these relations with those from Italy [Sabetta
and Pugliese, 1996] and from active tectonic regions in Europe [Ambraseys et al., 1996]. Nonetheless,
caution should be used when using these relations in areas outside of WNA.
Four attenuation relations commonly used to develop engineering estimates of strong ground motion
in WNA are presented [Abrahamson and Silva, 1997; Boore et al., 1997; Campbell and Bozorgnia, in press;
Sadigh et al., 1993, 1997]. All of these relations are empirically based, meaning that they were developed
from a regression analysis of strong-motion recordings. The Campbell and Bozorgnia attenuation rela-
tion, published in an abbreviated and slightly different form in Campbell and Bozorgnia [2000], is
included although it is relatively new, because the authors recommend that it supersede that of Campbell
[1997, 2000c, 2001] and is being used in place of this latter relation in the 2002 update of the seismic
hazard maps produced by the U.S. Geological Survey. This is noteworthy because many engineering
studies use the same attenuation relations used by the U.S. Geological Survey in order to be consistent
with the hazard defined in these maps.
5.6.1.1 The Abrahamson and Silva Model
The attenuation relation of Abrahamson and Silva [1997] can be represented by the expression:
( ) ( )
ln Y = f1 MW , rrup + f 2 ( MW ) F + f3 MW , rrup HW + f 4 (S, ARock ) (5.43)
c + c ( M − 6.4) + c 3 (8.5 − M W )
c4
1 2 W
(
f 1 M W , rrup )
=
[ ]
+ c 5 + c 6 ( M W − 6.4) ln R for M W ≤ 6.4
(5.44)
c 1 + c 7 ( M W − 6.4) + c 3 (8.5 − M W )
c4
[ ]
+ c 5 + c 6 ( M W − 6.4) ln R for M W > 6.4
R = rrup
2
+ c 82 (5.45)
c 9 for M W ≤ 5.8
(c − c ) ( MW − 5.8)
f 2 ( M W ) = c 9 + 10 9 for 5.8 < M W < 6.4 (5.46)
0.6
c 10 for M W ≥ 6.4
( )
f 3 M W , rrup = f HW ( M W ) f HW rrup ( ) (5.47)
0 for M W ≤ 5.5
f HW ( M W ) = M W − 5.5 for 5.5 < M W < 6.5 (5.48)
1 for M W ≥ 6.5
0 for rrup ≤ 4 km
r −4
c rup for 4 < rrup ≤ 8 km
11 4
( )
f HW rrup
= c 11
rrup − 18
for 8 < rrup ≤ 18 km (5.49)
[ ]
f 4 ( S, A Rock ) = c 12 + c 13 ln ( A Rock + c 14 ) SSoil (5.50)
The parameter Arock is the value of the average horizontal or vertical component of PGA for generic
rock. The indicator variable HW quantifies the effect of the hanging wall.
The standard deviation of lnY is separated into its inter- and intra-earthquake components. However,
to be consistent with engineering convention, only the total value is reported by the authors. The total
standard deviation is defined as a function of magnitude according to the expression:
c 15 for M W ≤ 5.0
σ lnY ( M W ) = c15 − c 16 ( M W − 5) for 5.0 < M W < 7.0 (5.51)
c − 2c for M W ≥ 7.0
15 16
The period-independent regression coefficients in the above equations are c2 = 0.512, c4 = 2, c6 = 0.17,
c7 = −0.144, and c14 = 0.03 for the average horizontal component of ground motion, and c2 = 0.909,
c4 = 3, c6 = 0.06, c7 = 0.275, and c14 = 0.3 for the vertical component of ground motion. The remaining
regression coefficients are listed in Tables 5.6 and 5.7.
The relation predicts ground motion for generic rock, equivalent to the condition SRock = 1, when
SSoil = 0. When SSoil = 1, it predicts ground motion for generic soil (see Table 5.4). The expression for local
site conditions accounts for the nonlinear behavior of soil by linking soil amplification to the amplitude
of PGA on rock. The relation includes soils up to 20 m thick in the definition of rock. This thickness is
greater than that used in most other relations, making the generic rock classification softer than that
defined by Boore and Joyner [1997]. One of the authors [Walt Silva, personal communication, 2001]
recommends using a 30-m velocity of 290 m/sec for generic soil and 510 m/sec for generic rock when
using this attenuation relation. Nonetheless, it is generally assumed that the rock classification used in
this relation is approximately the same as that used in other relations when making engineering estimates
of ground motion in WNA (Table 5.4). Faulting mechanism is defined as reverse, reverse-oblique, and
TABLE 5.6 Coefficients for Abrahamson and Silva Attenuation Relation: Horizontal Component
Tn (s) c1 c3 c5 c8 c9 c10 c11 c12 c13 c15 c16
PGA 1.640 0.0000 –1.1450 5.60 0.610 0.260 0.370 –0.417 –0.230 0.70 0.135
0.02 1.640 0.0000 –1.1450 5.60 0.610 0.260 0.370 –0.417 –0.230 0.70 0.135
0.03 1.690 0.0143 –1.1450 5.60 0.610 0.260 0.370 –0.470 –0.230 0.70 0.135
0.04 1.780 0.0245 –1.1450 5.60 0.610 0.260 0.370 –0.555 –0.251 0.71 0.135
0.05 1.870 0.0280 –1.1450 5.60 0.610 0.260 0.370 –0.620 –0.267 0.71 0.135
0.06 1.940 0.0300 –1.1450 5.60 0.610 0.260 0.370 –0.665 –0.280 0.72 0.135
0.075 2.037 0.0300 –1.1450 5.58 0.610 0.260 0.370 –0.628 –0.280 0.73 0.135
0.09 2.100 0.0300 –1.1450 5.54 0.610 0.260 0.370 –0.609 –0.280 0.74 0.135
0.10 2.160 0.0280 –1.1450 5.50 0.610 0.260 0.370 –0.598 –0.280 0.74 0.135
0.12 2.272 0.0180 –1.1450 5.39 0.610 0.260 0.370 –0.591 –0.280 0.75 0.135
0.15 2.407 0.0050 –1.1450 5.27 0.610 0.260 0.370 –0.577 –0.280 0.75 0.135
0.17 2.430 –0.0040 –1.1350 5.19 0.610 0.260 0.370 –0.522 –0.265 0.76 0.135
0.20 2.406 –0.0138 –1.1150 5.10 0.610 0.260 0.370 –0.445 –0.245 0.77 0.135
0.24 2.293 –0.0238 –1.0790 4.97 0.610 0.232 0.370 –0.350 –0.223 0.77 0.135
0.30 2.114 –0.0360 –1.0350 4.80 0.610 0.198 0.370 –0.219 –0.195 0.78 0.135
0.36 1.955 –0.0460 –1.0052 4.62 0.610 0.170 0.370 –0.123 –0.173 0.79 0.135
0.40 1.860 –0.0518 –0.9880 4.52 0.610 0.154 0.370 –0.065 –0.160 0.79 0.135
0.46 1.717 –0.0594 –0.9652 4.38 0.592 0.132 0.370 0.020 –0.136 0.80 0.132
0.50 1.615 –0.0635 –0.9515 4.30 0.581 0.119 0.370 0.085 –0.121 0.80 0.130
0.60 1.428 –0.0740 –0.9218 4.12 0.557 0.091 0.370 0.194 –0.089 0.81 0.127
0.75 1.160 –0.0862 –0.8852 3.90 0.528 0.057 0.331 0.320 –0.050 0.81 0.123
0.85 1.020 –0.0927 –0.8648 3.81 0.512 0.038 0.309 0.370 –0.028 0.82 0.121
1.00 0.828 –0.1020 –0.8383 3.70 0.490 0.013 0.281 0.423 0.000 0.83 0.118
1.50 0.260 –0.1200 –0.7721 3.55 0.438 –0.049 0.210 0.600 0.040 0.84 0.110
2.00 –0.150 –0.1400 –0.7250 3.50 0.400 –0.094 0.160 0.610 0.040 0.85 0.105
3.00 –0.690 –0.1726 –0.7250 3.50 0.400 –0.156 0.089 0.630 0.040 0.87 0.097
4.00 –1.130 –0.1956 –0.7250 3.50 0.400 –0.200 0.039 0.640 0.040 0.88 0.092
5.00 –1.460 –0.2150 –0.7250 3.50 0.400 –0.200 0.000 0.664 0.040 0.89 0.087
Source: Adapted from Abrahamson, N.A. and Silva, W.J. 1997. “Empirical Response Spectral Attenuation Relations for
Shallow Crustal Earthquakes,” Seismol. Res. Lett., 68, 94–127.
other than reverse and reverse-oblique (i.e., strike slip and normal) without specifically defining these
categories in terms of rake angle. As a result, it is somewhat ambiguous how the faulting mechanism
parameter F should be determined for a specific fault. Some guidance can be obtained from the relationship
between F and λ used in other models (Table 5.5).
5.6.1.2 The Boore et al. Model
The attenuation relation of Boore et al. [1997] can be represented by the expression:
( )
ln Y = f 1 M W , r jb + f 2 (VS 30 ) (5.52)
( )
f 1 M W , r jb = c 1 + c 2 ( M W − 6) + c 3 ( M W − 6) + c 4 ln R
2
(5.53)
R = r jb2 + c 52 (5.54)
TABLE 5.7 Coefficients for Abrahamson and Silva Attenuation Relation: Vertical Component
Tn (s) c1 c3 c5 c8 c9 c10 c11 c12 c13 c15 c16
PGA 1.642 0.0000 –1.2520 6.00 0.390 –0.050 0.630 –0.140 –0.220 0.76 0.085
0.02 1.642 0.0000 –1.2520 6.00 0.390 –0.050 0.630 –0.140 –0.220 0.76 0.085
0.03 2.100 0.0000 –1.3168 6.00 0.432 –0.050 0.630 –0.140 –0.220 0.76 0.085
0.04 2.420 0.0000 –1.3700 6.00 0.469 –0.050 0.630 –0.140 –0.220 0.76 0.085
0.05 2.620 –0.0002 –1.3700 6.00 0.496 –0.050 0.630 –0.140 –0.220 0.76 0.085
0.06 2.710 –0.0004 –1.3700 6.00 0.518 –0.050 0.630 –0.140 –0.220 0.76 0.085
0.075 2.750 –0.0007 –1.3700 6.00 0.545 –0.050 0.630 –0.129 –0.220 0.76 0.085
0.09 2.730 –0.0009 –1.3700 6.00 0.567 –0.050 0.630 –0.119 –0.220 0.76 0.085
0.10 2.700 –0.0010 –1.3700 6.00 0.580 –0.050 0.630 –0.114 –0.220 0.76 0.085
0.12 2.480 –0.0015 –1.2986 6.00 0.580 –0.017 0.630 –0.104 –0.220 0.74 0.075
0.15 2.170 –0.0022 –1.2113 6.00 0.580 0.024 0.630 –0.093 –0.220 0.72 0.063
0.17 1.960 –0.0025 –1.1623 5.72 0.580 0.047 0.604 –0.087 –0.220 0.70 0.056
0.20 1.648 –0.0030 –1.0987 5.35 0.580 0.076 0.571 –0.078 –0.220 0.69 0.050
0.24 1.312 –0.0035 –1.0274 4.93 0.580 0.109 0.533 –0.069 –0.220 0.69 0.050
0.30 0.878 –0.0042 –0.9400 4.42 0.580 0.150 0.488 –0.057 –0.220 0.69 0.050
0.36 0.617 –0.0047 –0.9004 4.01 0.571 0.150 0.450 –0.048 –0.220 0.69 0.050
0.40 0.478 –0.0050 –0.8776 3.77 0.539 0.150 0.428 –0.043 –0.220 0.69 0.050
0.46 0.271 –0.0056 –0.8472 3.45 0.497 0.150 0.400 –0.035 –0.220 0.69 0.050
0.50 0.145 –0.0060 –0.8291 3.26 0.471 0.150 0.383 –0.031 –0.220 0.69 0.050
0.60 –0.087 –0.0068 –0.7896 2.85 0.416 0.150 0.345 –0.022 –0.220 0.69 0.050
0.75 –0.344 –0.0083 –0.7488 2.50 0.348 0.150 0.299 –0.010 –0.220 0.69 0.050
0.85 –0.469 –0.0097 –0.7451 2.50 0.309 0.150 0.273 –0.004 –0.220 0.69 0.050
1.00 –0.602 –0.0115 –0.7404 2.50 0.260 0.150 0.240 0.004 –0.220 0.69 0.050
1.50 –0.966 –0.0180 –0.7285 2.50 0.260 0.058 0.240 0.025 –0.220 0.69 0.050
2.00 –1.224 –0.0240 –0.7200 2.50 0.260 –0.008 0.240 0.040 –0.220 0.69 0.050
3.00 –1.581 –0.0431 –0.7200 2.50 0.260 –0.100 0.240 0.040 –0.220 0.72 0.050
4.00 –1.857 –0.0565 –0.7200 2.50 0.260 –0.100 0.240 0.040 –0.220 0.75 0.050
5.00 –2.053 –0.0670 –0.7200 2.50 0.260 –0.100 0.240 0.040 –0.220 0.78 0.050
Source: Adapted from Abrahamson, N.A. and Silva, W.J. 1997. “Empirical Response Spectral Attenuation Relations for
Shallow Crustal Earthquakes,” Seismol. Res. Lett., 68, 94–127.
The standard deviation of lnY is separated into its inter- and intra-earthquake components. Further-
more, the authors report the total standard deviation, which includes the randomness between the two
horizontal components of ground motion, or what constitutes the random rather than the average
horizontal component. To be consistent with standard engineering convention, the total standard devi-
ation excluding the component-to-component randomness is reported here. The regression coefficients
and the total standard deviation are listed in Table 5.8. The relation is considered valid for MW = 5.5 to
7.5 and rrup ≤ 80 km.
The relation accounts for site response through the 30-m velocity VS30. Some typical values for VS30
are provided in Tables 5.2 and 5.4 [see also Wills and Silva, 1998]. This parameter must be set to some
value in order for the prediction to be valid. Boore and Joyner [1997] recommend using VS30 = 310 m/sec
and 620 m/sec for generic soil and generic rock, respectively. The relationship between the faulting
categories and the rake angle λ is given in Table 5.5.
5.6.1.3 The Campbell and Bozorgnia Model
The attenuation relation of Campbell and Bozorgnia [in press] can be represented by the expression:
PGA –0.242 –0.313 –0.117 0.527 0.000 –0.778 5.57 –0.371 1396 0.468
0.10 1.059 1.006 1.087 0.753 –0.226 –0.934 6.27 –0.212 1112 0.440
0.11 1.130 1.072 1.164 0.732 –0.230 –0.937 6.65 –0.211 1291 0.437
0.12 1.174 1.109 1.215 0.721 –0.233 –0.939 6.91 –0.215 1452 0.437
0.13 1.200 1.128 1.246 0.711 –0.233 –0.939 7.08 –0.221 1596 0.435
0.14 1.208 1.135 1.261 0.707 –0.230 –0.938 7.18 –0.228 1718 0.435
0.15 1.204 1.128 1.264 0.702 –0.228 –0.937 7.23 –0.238 1820 0.435
0.16 1.192 1.112 1.257 0.702 –0.226 –0.935 7.24 –0.248 1910 0.436
0.17 1.173 1.090 1.242 0.702 –0.221 –0.933 7.21 –0.258 1977 0.436
0.18 1.151 1.063 1.222 0.705 –0.216 –0.930 7.16 –0.270 2037 0.435
0.19 1.122 1.032 1.198 0.709 –0.212 –0.927 7.10 –0.281 2080 0.435
0.20 1.089 0.999 1.170 0.711 –0.207 –0.924 7.02 –0.292 2118 0.435
0.22 1.019 0.925 1.104 0.721 –0.198 –0.918 6.83 –0.315 2158 0.438
0.24 0.941 0.847 1.033 0.732 –0.189 –0.912 6.62 –0.338 2178 0.439
0.26 0.861 0.764 0.958 0.744 –0.180 –0.906 6.39 –0.360 2173 0.439
0.28 0.780 0.681 0.881 0.758 –0.168 –0.899 6.17 –0.381 2158 0.441
0.30 0.700 0.598 0.803 0.769 –0.161 –0.893 5.94 –0.401 2133 0.443
0.32 0.619 0.518 0.725 0.783 –0.152 –0.888 5.72 –0.420 2104 0.445
0.34 0.540 0.439 0.648 0.794 –0.143 –0.882 5.50 –0.438 2070 0.449
0.36 0.462 0.361 0.570 0.806 –0.136 –0.877 5.30 –0.456 2032 0.450
0.38 0.385 0.286 0.495 0.820 –0.127 –0.872 5.10 –0.472 1995 0.453
0.40 0.311 0.212 0.423 0.831 –0.120 –0.867 4.91 –0.487 1954 0.454
0.42 0.239 0.140 0.352 0.840 –0.113 –0.862 4.74 –0.502 1919 0.458
0.44 0.169 0.073 0.282 0.852 –0.108 –0.858 4.57 –0.516 1884 0.460
0.46 0.102 0.005 0.217 0.863 –0.101 –0.854 4.41 –0.529 1849 0.463
0.48 0.036 –0.058 0.151 0.873 –0.097 –0.850 4.26 –0.541 1816 0.464
0.50 –0.025 –0.122 0.087 0.884 –0.090 –0.846 4.13 –0.553 1782 0.469
0.55 –0.176 –0.268 –0.063 0.907 –0.078 –0.837 3.82 –0.579 1710 0.474
0.60 –0.314 –0.401 –0.203 0.928 –0.069 –0.830 3.57 –0.602 1644 0.480
0.65 –0.440 –0.523 –0.331 0.946 –0.060 –0.823 3.36 –0.622 1592 0.485
0.70 –0.555 –0.634 –0.452 0.962 –0.053 –0.818 3.20 –0.639 1545 0.492
0.75 –0.661 –0.737 –0.562 0.979 –0.046 –0.813 3.07 –0.653 1507 0.497
0.80 –0.760 –0.829 –0.666 0.992 –0.041 –0.809 2.98 –0.666 1476 0.502
0.85 –0.851 –0.915 –0.761 1.006 –0.037 –0.805 2.92 –0.676 1452 0.505
0.90 –0.933 –0.993 –0.848 1.018 –0.035 –0.802 2.89 –0.685 1432 0.511
0.95 –1.010 –1.066 –0.932 1.027 –0.032 –0.800 2.88 –0.692 1416 0.515
1.00 –1.080 –1.133 –1.009 1.036 –0.032 –0.798 2.90 –0.698 1406 0.520
1.10 –1.208 –1.249 –1.145 1.052 –0.030 –0.795 2.99 –0.706 1396 0.528
1.20 –1.315 –1.345 –1.265 1.064 –0.032 –0.794 3.14 –0.710 1400 0.533
1.30 –1.407 –1.428 –1.370 1.073 –0.035 –0.793 3.36 –0.711 1416 0.540
1.40 –1.483 –1.495 –1.460 1.080 –0.039 –0.794 3.62 –0.709 1442 0.545
1.50 –1.550 –1.552 –1.538 1.085 –0.044 –0.796 3.92 –0.704 1479 0.550
1.60 –1.605 –1.598 –1.608 1.087 –0.051 –0.798 4.26 –0.697 1524 0.554
1.70 –1.652 –1.634 –1.668 1.089 –0.058 –0.801 4.62 –0.689 1581 0.558
1.80 –1.689 –1.663 –1.718 1.087 –0.067 –0.804 5.01 –0.679 1644 0.561
1.90 –1.720 –1.685 –1.763 1.087 –0.074 –0.808 5.42 –0.667 1714 0.565
2.00 –1.743 –1.699 –1.801 1.085 –0.085 –0.812 5.85 –0.655 1795 0.566
Source: Adapted from Boore, D.M., Joyner, W.B., and Fumal, T.E. 1997. “Equations for Estimating Horizontal Response
Spectra and Peak Acceleration from Western North American Earthquakes: A Summary of Recent Work,” Seismol. Res. Lett.,
68, 128–153.
f 1 ( M W ) = c 2 M W + c 3 (8.5 − MW )
2
(5.58)
[ ]
2
f 2 ( M W , rseis , S) = rseis + g ( S) exp c 8 M W + c 9 (8.5 − M W )
2 2 2
(5.59a)
The parameter HW quantifies the effect of the hanging wall and will always evaluate to zero for firm
soil and for a horizontal distance of 5 km or greater from the rupture plane. This distinguishes it from
the hanging-wall factor of Abrahamson and Silva [1997].
The standard deviation of lnY is defined as a function of magnitude according to the expression:
where PGA is either uncorrected PGA or corrected PGA, depending on the application (see footnote
to Table 5.9). The regression coefficients are listed in Tables 5.9 and 5.10. The relation is considered
valid for MW ≥ 4.7 and rseis ≤ 60 km.
The relation predicts ground motion for firm soil, equivalent to the condition SFS = 1, unless one of
the site parameters in g (S) and f4 (S) is set to one, in which case it predicts ground motion for either
very firm soil, soft rock, or firm rock. Further guidance on evaluating the relation for various site
conditions is given tin Table 5.4. The relationship between the faulting mechanism parameters FRV (reverse
faulting with dip greater than 45°) and FTH (thrust faulting with dip less than or equal to 45°) and the
rake angle λ is given in Table 5.5. Sediment depth D was evaluated and found to be important, but it
was not included as a parameter, since it is rarely used in engineering practice. If desired, sediment depth
can be included in an estimate of ground motion by using the attenuation relation developed by Campbell
[1997, 2000c, 2001].
5.6.1.4 The Sadigh et al. Model
The attenuation relation of Sadigh et al. [1993, 1997] can be represented by the expression:
(
ln Y = c1F + f MW , rrup ) (5.64)
( ) (
f MW , rrup = c 2 + c 3 MW + c 4 (8.5 − MW ) + c 5 ln R + c 6 ln rrup + 2
2.5
) (5.65)
c 9 for MW ≤ c13
σ lnY = c10 − c11MW for c13 < MW < c14 (5.67)
c for MW ≥ c14
12
The relation can be used to estimate ground motion for generic soil, equivalent to the condition
SSoil = 1, or generic rock, equivalent to the condition SRock = 1 (Table 5.4). The regression coefficients for
generic rock are listed in Tables 5.11 and 5.12 and those for generic soil in Table 5.13. No regression
coefficients are given for the vertical component of ground motion for soil. The relation is considered
valid for MW = 4.0 to 8.0 and rrup ≤ 100 km.
The authors say that they define rock as a deposit with less than a meter of soil. However, one of the
authors [Bob Youngs, personal communication, 2001] says that this goal was not actually achieved and
that the rock and soil categories are generally consistent with those of Abrahamson and Silva [1997].
Youngs therefore recommends using VS30 = 290 m/sec and 510 m/sec to represent generic soil and generic
rock in the Sadigh et al. attenuation relations. Nonetheless, it is generally assumed that the rock classi-
fication used in this relation is approximately the same as that used in other relations when making
engineering estimates of ground motion in WNA (Table 5.4). The relationship between the faulting
mechanism parameter F and the rake angle λ is given in Table 5.5.
Unc PGA –2.896 0.812 0.000 –1.318 0.187 –0.029 –0.064 0.616 0 0.179 0.307 –0.062 –0.195 –0.320 0.370 0.964 0.263
Cor PGA –4.033 0.812 0.036 –1.061 0.041 –0.005 –0.018 0.766 0.034 0.343 0.351 –0.123 –0.138 –0.289 0.370 0.920 0.219
0.05 –3.740 0.812 0.036 –1.121 0.058 –0.004 –0.028 0.724 0.032 0.302 0.362 –0.140 –0.158 –0.205 0.370 0.940 0.239
0.075 –3.076 0.812 0.050 –1.252 0.121 –0.005 –0.051 0.648 0.040 0.243 0.333 –0.150 –0.196 –0.208 0.370 0.952 0.251
0.10 –2.661 0.812 0.060 –1.308 0.166 –0.009 –0.068 0.621 0.046 0.224 0.313 –0.146 –0.253 –0.258 0.370 0.958 0.257
0.15 –2.270 0.812 0.041 –1.324 0.212 –0.033 –0.081 0.613 0.031 0.318 0.344 –0.176 –0.267 –0.284 0.370 0.974 0.273
0.20 –2.771 0.812 0.030 –1.153 0.098 –0.014 –0.038 0.704 0.026 0.296 0.342 –0.148 –0.183 –0.359 0.370 0.981 0.280
0.30 –2.999 0.812 0.007 –1.080 0.059 –0.007 –0.022 0.752 0.007 0.359 0.385 –0.162 –0.157 –0.585 0.370 0.984 0.283
0.40 –3.511 0.812 –0.015 –0.964 0.024 –0.002 –0.005 0.842 –0.016 0.379 0.438 –0.078 –0.129 –0.557 0.370 0.987 0.286
0.50 –3.556 0.812 –0.035 –0.964 0.023 –0.002 –0.004 0.842 –0.036 0.406 0.479 –0.122 –0.130 –0.701 0.370 0.990 0.289
0.75 –3.709 0.812 –0.071 –0.964 0.021 –0.002 –0.002 0.842 –0.074 0.347 0.419 –0.108 –0.124 –0.796 0.331 1.021 0.320
1.0 –3.867 0.812 –0.101 –0.964 0.019 0 0 0.842 –0.105 0.329 0.338 –0.073 –0.072 –0.858 0.281 1.021 0.320
1.5 –4.093 0.812 –0.150 –0.964 0.019 0 0 0.842 –0.155 0.217 0.188 –0.079 –0.056 –0.954 0.210 1.021 0.320
2.0 –4.311 0.812 –0.180 –0.964 0.019 0 0 0.842 –0.187 0.060 0.064 –0.124 –0.116 –0.916 0.160 1.021 0.320
3.0 –4.817 0.812 –0.193 –0.964 0.019 0 0 0.842 –0.200 –0.079 0.021 –0.154 –0.117 –0.873 0.089 1.021 0.320
4.0 –5.211 0.812 –0.202 –0.964 0.019 0 0 0.842 –0.209 –0.061 0.057 –0.054 –0.261 –0.889 0.039 1.021 0.320
Note: Uncorrected PGA is to be used only when an estimate of PGA is required. Corrected PGA is to be used when an estimate of PGA compatible with PSA is required.
Source: Adapted from Campbell, K.W. and Bozorgnia, Y., in press. “Updated Near-Source Ground Motion (Attenuation) Relations for the Horizontal and Vertical Components of Peak
Ground Acceleration and Acceleration Response Spectra,” Bull. Seismol. Soc. Am.
5-33
© 2003 by CRC Press LLC
0068_C05_fm Page 34 Thursday, August 1, 2002 7:59 AM
5-34
TABLE 5.10 Coefficients for Campbell and Bozorgnia Attenuation Relation: Vertical Component
Tn (s) c1 c2 c3 c4 c5 c6 c7 c8 c9 c10 c11 c12 c13 c14 c15 c17 c17
Unc PGA –2.807 0.756 0 –1.391 0.191 0.044 –0.014 0.544 0 0.091 0.223 –0.096 –0.212 –0.199 0.630 1.003 0.320
Cor PGA –3.108 0.756 0 –1.287 0.142 0.046 –0.040 0.587 0 0.253 0.173 –0.135 –0.138 –0.256 0.630 0.975 0.274
0.05 –1.918 0.756 0 –1.517 0.309 0.069 –0.023 0.498 0 0.058 0.100 –0.195 –0.274 –0.219 0.630 1.031 0.330
0.075 –1.504 0.756 0 –1.551 0.343 0.083 0.000 0.487 0 0.135 0.182 –0.224 –0.303 –0.263 0.630 1.031 0.330
0.10 –1.672 0.756 0 –1.473 0.282 0.062 0.001 0.513 0 0.168 0.210 –0.198 –0.275 –0.252 0.630 1.031 0.330
0.15 –2.323 0.756 0 –1.280 0.171 0.045 0.008 0.591 0 0.223 0.238 –0.170 –0.175 –0.270 0.630 1.031 0.330
0.20 –2.998 0.756 0 –1.131 0.089 0.028 0.004 0.668 0 0.234 0.256 –0.098 –0.041 –0.311 0.571 1.031 0.330
0.30 –3.721 0.756 0.007 –1.028 0.050 0.010 0.004 0.736 0.007 0.249 0.328 –0.026 0.082 –0.265 0.488 1.031 0.330
0.40 –4.536 0.756 –0.015 –0.812 0.012 0 0 0.931 –0.018 0.299 0.317 –0.017 0.022 –0.257 0.428 1.031 0.330
0.50 –4.651 0.756 –0.035 –0.812 0.012 0 0 0.931 –0.043 0.243 0.354 –0.020 0.092 –0.293 0.383 1.031 0.330
0.75 –4.903 0.756 –0.071 –0.812 0.012 0 0 0.931 –0.087 0.295 0.418 0.078 0.091 –0.349 0.299 1.031 0.330
1.0 –4.950 0.756 –0.101 –0.812 0.012 0 0 0.931 –0.124 0.266 0.315 0.043 0.101 –0.481 0.240 1.031 0.330
1.5 –5.073 0.756 –0.150 –0.812 0.012 0 0 0.931 –0.184 0.171 0.211 –0.038 –0.018 –0.518 0.240 1.031 0.330
2.0 –5.292 0.756 –0.180 –0.812 0.012 0 0 0.931 –0.222 0.114 0.115 0.033 –0.022 –0.503 0.240 1.031 0.330
Note: Uncorrected PGA is to be used only when an estimate of PGA is required. Corrected PGA is to be used when an estimate of PGA compatible with PSA is required.
Source: Adapted from Campbell, K.W. and Bozorgnia, Y. in press. “Updated Near-Source Ground Motion (Attenuation) Relations for the Horizontal and Vertical Components of Peak
Ground Acceleration and Acceleration Response Spectra,” Bull. Seismol. Soc. Am.
analysis o f seismolo gical-base d est imates of strong g round mot ion. The Campbell model was
developed using the h ybrid empirical me thod, in which attenuation relations in WNA were modified
for use in EN A using se ismolo gical est imates of ground mot ion in WNA and ENA. Because o f its
use o f empirical att enuation relations in WNA, this latt er model inherently incorporates finite-
fault ing effects. It should b e noted that the re are two additional attenuation relations [Frankel et
al., 1996; Somerville et al., 2001] that ha ve not b een publishe d in an international jour nal but which
are being use d in the 2002 updat e of the se ismic hazar d maps p roduced by the U.S. Geolo gical
Survey (the F rankel et al. relation was also use d in the 1996 v ersion of these hazar d maps). This is
noteworthy because man y engineering st udies a dopt the att enuation relations use d by the U.S.
Geolo gical Survey in order to be consist ent with the hazar d defined in these maps .
The three selected attenuation relations provide ground-motion estimates for a very hard rock, typ-
ically referred to as ENA hard rock [Boore and Joyner, 1997; see also Table 5.4]. ENA hard rock, with a
30-m velocity of around 2800 m/sec, is not a typical site condition encountered in engineering investi-
gations, even in ENA, so some adjustment must be made for local site conditions. Atkinson and Boore
[1995, 1997] provide a site term to use with their attenuation relation but this term corresponds to a
very firm soil or soft rock and is calculated using the stochastic method assuming linear soil response.
As a result, this site term may not be appropriate for the larger ground motions of engineering interest.
One simple approach is to assume that ENA hard rock represents building-code Site Class A and use the
site factors in the U.S. building codes to approximate the effect of local soil conditions for other site
classes. One of the problems with this approach is that these soil factors are given only for spectral
accelerations averaged over short periods (Tn = 0.1 to 0.5 sec) or long periods (Tn = 0.5 to 2.0 sec), so
one would have to use judgment in order to use these factors for individual periods. Another problem
is that Site Class A is usually associated with a 30-m velocity of around 1890 m/sec [Savy et al., 1987],
which is much less than that for ENA hard rock. Boore [in press] suggests that nonlinear soil effects should
be taken into account using a nonlinear site-response analysis. Additional guidance on incorporating site
effects is provided by Savy and coworkers [1987], EPRI [1993], and Hwang and coworkers [1997].
5.6.2.1 The Atkinson and Boore Model
The attenuation relation of Atkinson and Boore [1995, 1997] can be represented by the expression:
( )
ln Y = f 1 M W , rhypo + f 2 ( S) (5.68)
( )
f 1 M W , rhypo = c 1 + c 2 ( M W − 6) + c 3 ( M W − 6) − ln rhypo + c 4 rhypo
2
(5.69)
f 2 ( S) = c 5 SDeep (5.70)
The regression coefficients and the recommended values of the standard deviation of ln Y [taken from
Atkinson, 1995] are listed in Table 5.14. The regression coefficient c5 was theoretically developed using
Equation 5.30, so caution should be used when applying this coefficient to very high ground-motion
amplitudes where nonlinear soil behavior may be expected. The relation is considered valid for MW =
4.0 to 7.25 and rhypo = 10 to 500 km. The attenuation relations were developed as an approximation to
the simulated ground-motion values for use with PSHA. The relations overestimate the simulated values
for MW ≤ 5.5 and rhypo > 30 km. This bias is not an issue for most seismic hazard studies. However, a
more accurate estimate of ground motion can be obtained from the tables provided in the appendices
of Atkinson and Boore [1995, 1997].
The relation can be used to estimate ground motion for ENA hard rock, equivalent to the condition
SENA = 1, or ENA deep stiff soil, equivalent to setting SDeep = 1 (see also Table 5.4). A deep stiff soil is
defined as a very stiff site with soil depth > 60 m and VS30 ≈ 500 m/sec. This is similar to the average
30-m velocity for building-code Site Class C (Tables 5.2 and 5.4) but its site-response characteristics are
MW ≤ 6.5
PGA 0.182 –0.624 1.0 0 –2.100 0 3.6564 0.250 0 1.39 0.14 0.38 0 7.21
0.05 0.182 –0.090 1.0 0.006 –2.128 –0.082 3.6564 0.250 0 1.39 0.14 0.38 0 7.21
0.07 0.182 0.110 1.0 0.006 –2.128 –0.082 3.6564 0.250 0 1.40 0.14 0.39 0 7.21
0.09 0.182 0.212 1.0 0.006 –2.140 –0.052 3.6564 0.250 0 1.40 0.14 0.39 0 7.21
0.10 0.182 0.275 1.0 0.006 –2.148 –0.041 3.6564 0.250 0 1.41 0.14 0.40 0 7.21
0.12 0.182 0.348 1.0 0.005 –2.162 –0.014 3.6564 0.250 0 1.41 0.14 0.40 0 7.21
0.14 0.182 0.307 1.0 0.004 –2.144 0 3.6564 0.250 0 1.42 0.14 0.41 0 7.21
0.15 0.182 0.285 1.0 0.002 –2.130 0 3.6564 0.250 0 1.42 0.14 0.41 0 7.21
0.17 0.182 0.239 1.0 0 –2.110 0 3.6564 0.250 0 1.42 0.14 0.41 0 7.21
MW ≤ 6.5
PGA 0.0953 –0.4300 1.0 0 –2.300 0 3.5701 0.228 0.68 3.08 0.40 0.48 6.0 6.5
0.04 0.0953 0.3379 1.0 0 –2.450 0 3.5701 0.228 0.75 2.91 0.36 0.57 6.0 6.5
0.05 0.0953 0.5041 1.0 0 –2.450 0 3.5701 0.228 0.75 2.91 0.36 0.57 6.0 6.5
0.06 0.0953 0.6095 1.0 0 –2.450 0 3.5701 0.228 0.75 2.91 0.36 0.57 6.0 6.5
0.07 0.0953 0.6896 1.0 0 –2.450 0 3.5701 0.228 0.75 2.91 0.36 0.57 6.0 6.5
0.09 0.0953 0.6718 1.0 –0.00330 –2.420 0 3.5701 0.228 0.75 2.91 0.36 0.57 6.0 6.5
0.10 0.0953 0.6252 1.0 –0.00468 –2.400 0 3.5701 0.228 0.75 2.91 0.36 0.57 6.0 6.5
0.12 0.0953 0.5535 1.0 –0.00707 –2.380 0 3.5701 0.228 0.75 2.91 0.36 0.57 6.0 6.5
0.14 0.0953 0.3813 1.0 –0.00909 –2.333 0 3.5701 0.228 0.75 2.91 0.36 0.57 6.0 6.5
Source: Adapted from Sadigh, K. et al. 1993. “Specification of Long-Period Ground Motions: Updated Attenuation Relationships for Rock Site Conditions
and Adjustment Factors for Near-Fault Effects,” in Proc. ATC-17-1 Seminar on Seismic Isolation, Passive Energy Dissipation, and Active Control, Vol. I, March
11–12, San Francisco, pp. 11–23, Applied Technology Council, Redwood City, CA.
5-39
© 2003 by CRC Press LLC
0068_C05_fm Page 40 Thursday, August 1, 2002 7:59 AM
5-40
TABLE 5.13 Coefficients for Sadigh et al. Soil Attenuation Relation
Tn (s) c1 c2 c3 c4 c5 c6 c7 c8 c9 c10 c11 c12 c13 c14
MW ≤ 6.5
PGA 0.2500 –2.1700 1.0 0 –1.700 0 2.1863 0.320 0 1.520 0.16 0.40 0 7.0
0.075 0.2500 –1.7128 1.0 0.005 –1.700 0 2.1863 0.320 0 1.540 0.16 0.42 0 7.0
0.10 0.2500 –1.5305 1.0 0.005 –1.700 0 2.1863 0.320 0 1.540 0.16 0.42 0 7.0
0.20 0.2500 –1.2513 1.0 –0.004 –1.700 0 2.1863 0.320 0 1.565 0.16 0.45 0 7.0
0.30 0.2500 –1.2153 1.0 –0.014 –1.700 0 2.1863 0.320 0 1.580 0.16 0.46 0 7.0
0.40 0.2254 –1.2449 1.0 –0.024 –1.700 0 2.1863 0.320 0 1.595 0.16 0.48 0 7.0
0.50 0.2291 –1.3206 1.0 –0.033 –1.700 0 2.1863 0.320 0 1.610 0.16 0.49 0 7.0
0.75 0.2292 –1.4690 1.0 –0.051 –1.700 0 2.1863 0.320 0 1.635 0.16 0.52 0 7.0
1.0 0.1910 –1.6035 1.0 –0.065 –1.700 0 2.1863 0.320 0 1.660 0.16 0.54 0 7.0
1.5 0.1480 –1.8465 1.0 –0.090 –1.700 0 2.1863 0.320 0 1.690 0.16 0.57 0 7.0
2.0 0.0973 –2.0699 1.0 –0.108 –1.700 0 2.1863 0.320 0 1.700 0.16 0.58 0 7.0
3.0 0.0396 –2.4501 1.0 –0.139 –1.700 0 2.1863 0.320 0 1.710 0.16 0.59 0 7.0
4.0 –0.0133 –2.7974 1.0 –0.160 –1.700 0 2.1863 0.320 0 1.710 0.16 0.59 0 7.0
MW > 6.5
PGA 0.2500 –2.1700 1.0 0 –1.700 0 0.3825 0.5882 0 1.520 0.16 0.40 0 7.0
0.075 0.2500 –1.7128 1.0 0.005 –1.700 0 0.3825 0.5882 0 1.540 0.16 0.42 0 7.0
0.10 0.2500 –1.5305 1.0 0.005 –1.700 0 0.3825 0.5882 0 1.540 0.16 0.42 0 7.0
0.20 0.2500 –1.2513 1.0 –0.004 –1.700 0 0.3825 0.5882 0 1.565 0.16 0.45 0 7.0
Source: Adapted from Sadigh, K. et al. 1997. “Attenuation Relationships for Shallow Crustal Earthquakes Based on California Strong Motion
Data,” Seismol. Res. Lett., 68, 180–189.
Source: Adapted from Atkinson, G.M. and Boore, D.M. 1995. “New Ground Motion Relations for Eastern
North America,” Bull. Seismol. Soc. Am., 85, 17–30; and Atkinson, G.M. and Boore, D.M. 1997. “Some
Comparisons between Recent Ground Motion Relations,” Seismol. Res. Lett., 68, 24–40.
likely to be different, because of the difference in the bedrock properties between ENA and WNA, the
source region for deriving the site-response characteristic of Site Class C used in the building codes.
If an estimate of ground motion for a given value of the magnitude measure mLg (mN in Canada) is
required, MW can be estimated from mLg, using the relation:
The use of hypocentral distance as the distance measure means that the relation is not strictly valid
for large earthquakes with extended rupture zones. Boore [in press] suggests using rhypo as an estimate
of the closest distance to fault rupture rrup in such cases. Guidance on incorporating local soil effects is
given in Section 5.6.2.
5.6.2.2 The Toro et al. Model
The attenuation relation of Toro et al. [1997] can be represented by the expression:
ln Y = c 1 + c 2 ( M − 6) + c 3 ( M − 6) + c 4 ln R + c 5 f ( R) + c 6 R
2
(5.72)
0 for R ≤ 100 km
f ( R) = (5.73)
ln ( R 100) for R > 100 km
R = r jb2 + c 72 (5.74)
Because the relations were developed from ground motions calculated using the stochastic method,
the standard deviation of lnY is theoretically rather than empirically defined from median estimates of
parametric uncertainty, according to the equation:
( )
σ lnY = σ ln2 Y ( M ) + σ ln2 Y r jb (5.75)
where σlnY (M) and σln Y(rjb) are computed by linear interpolation of the values given in Table 5.15.
Midcontinent Region, MW
PGA 2.20 0.81 0 –1.27 –0.11 –0.0021 9.3 0.55 0.59 0.50 0.54 0.20
0.029 4.00 0.79 0 –1.57 0.26 –0.0008 11.1 0.62 0.63 0.50 0.62 0.35
0.040 3.68 0.80 0 –1.46 0.31 –0.0013 10.5 0.62 0.63 0.50 0.57 0.29
0.10 2.37 0.81 0 –1.10 –0.08 –0.0040 8.3 0.59 0.61 0.50 0.50 0.17
0.20 1.73 0.84 0 –0.98 –0.32 –0.0042 7.5 0.60 0.64 0.56 0.45 0.12
Source: Adapted from Toro, G.R., Abrahamson, N.A., and Schneider, J.F. 1997. “Model of Strong Ground Motions from Earthquakes in Central and
Eastern North America: Best Estimates and Uncertainties,” Seismol. Res. Lett., 68, 41–57.
5-43
0068_C05_fm Page 44 Thursday, August 1, 2002 7:59 AM
The regression coefficients are listed in Table 5.15. There are two sets of regression coefficients, one
for the Midcontinent region and one for the Gulf Coast region. The Gulf Coast region represents the
lowlands that border the Gulf of Mexico and includes Louisiana, the southeastern coast of Texas, and
the southern areas of Alabama and Mississippi [EPRI, 1993]. The Midcontinent region covers the
remaining part of ENA. The authors provide no guidelines regarding the limitations of the relation.
However, some guidance can be obtained in EPRI [1993], the resource document for the relation, which
states that the relation was developed using the stochastic method for MW = 4.5 to 8.0 and rjb = 1 to 500 km.
The relation predicts ground motion for ENA hard rock, equivalent to the condition SENA = 1 (Table
5.4). The authors recommend that nonlinear soil effects should be modeled using the site factors given
in EPRI [1993], which were developed for sites of various soil thickness using dynamic site-response
analysis. Additional guidance on incorporating local soil effects is given in Section 5.6.2.
5.6.2.3 The Campbell Model
The attenuation relation of Campbell [in press] can be represented by the expression:
[ ( )] ( )
ln Y = c 1 + c 2 M W + c 3 (8.5 − M W ) + c 4 ln f 1 M W , rrup + f 2 rrup + (c 9 + c 10 M W ) rrup
2
(5.76)
( ) [
+ c 5 exp (c 6 M W ) ]
2
f 1 M W , rrup = rrup
2
(5.77)
0 for rrup ≤ 70 km
( )
f 2 rrup
(
= c7 ln rrup − ln 70 ) for 70 < rrup ≤ 130 km (5.78)
( ) (
c 7 ln rrup − ln 70 + c 8 ln rrup − ln 130 ) for rrup > 130 km
The standard deviation of lnY is defined as a function of magnitude according to the expression:
The regression coefficients are listed in Table 5.16. The relation is considered valid for MW ≥ 5.0 and
rrup ≥ 0 but it should be noted that the empirical database in WNA is restricted to earthquakes up to
magnitude 7.5. The author believes that it can be extrapolated to larger magnitudes because of its
physically based functional form and to large distances, beyond the 100-km limit of the WNA database,
because of its seismologically constrained attenuation rate. The relation can be used to estimate ground
motion for ENA hard rock, equivalent to the condition SENA = 1 (Table 5.4). Guidance on incorporating
local soil effects is given in Section 5.6.2.
5.6.3 Europe
There are many attenuation relations that have been developed for specific countries in Europe. The two
presented in this section [Ambraseys et al., 1996; Dahle et al., 1990] were chosen because they represent
the two major tectonic environments of the European continent as a whole rather than any one specific
country. Like North America, Europe can be divided into a shallow active tectonic region and a shallow
stable tectonic region. The attenuation characteristics of these two regions are similar to their counterparts
in North America. As a result, it is common to use the attenuation relations for WNA and ENA together
with European relations to develop engineering estimates of ground motion in Europe. Johnston [1994,
1996] and references therein show the geographic distribution of these two tectonic environments through-
out Europe and the rest of the world (Figure 5.6). Both of the selected relations were developed using the
empirical method, meaning that they are based on regression analysis of strong-motion recordings.
0.01 0.0305 0.633 –0.0427 –1.591 0.683 0.416 1.140 –0.873 –0.00428 0.000483 1.030 –0.0860 0.414
0.02 1.3535 0.630 –0.0404 –1.787 1.020 0.363 0.851 –0.715 –0.00388 0.000497 1.030 –0.0860 0.414
0.03 1.1860 0.622 –0.0362 –1.691 0.922 0.376 0.759 –0.922 –0.00367 0.000501 1.030 –0.0860 0.414
0.05 0.3736 0.616 –0.0353 –1.469 0.630 0.423 0.771 –1.239 –0.00378 0.000500 1.042 –0.0838 0.443
0.075 –0.0395 0.615 –0.0353 –1.383 0.491 0.463 0.955 –1.349 –0.00421 0.000486 1.052 –0.0838 0.453
0.10 –0.1475 0.613 –0.0353 –1.369 0.484 0.467 1.096 –1.284 –0.00454 0.000460 1.059 –0.0838 0.460
0.15 –0.1901 0.616 –0.0478 –1.368 0.461 0.478 1.239 –1.079 –0.00473 0.000393 1.068 –0.0838 0.469
0.20 –0.4328 0.617 –0.0586 –1.320 0.399 0.493 1.250 –0.928 –0.00460 0.000337 1.077 –0.0838 0.478
0.30 –0.6906 0.609 –0.0786 –1.280 0.349 0.502 1.241 –0.753 –0.00414 0.000263 1.081 –0.0838 0.482
0.50 –0.5907 0.534 –0.1379 –1.216 0.318 0.503 1.166 –0.606 –0.00341 0.000194 1.098 –0.0824 0.508
0.75 –0.5429 0.480 –0.1806 –1.184 0.304 0.504 1.110 –0.526 –0.00288 0.000160 1.105 –0.0806 0.528
1.0 –0.6104 0.451 –0.2090 –1.158 0.299 0.503 1.067 –0.482 –0.00255 0.000141 1.110 –0.0793 0.543
1.5 –0.9666 0.441 –0.2405 –1.135 0.304 0.500 1.029 –0.438 –0.00213 0.000119 1.099 –0.0771 0.547
2.0 –1.4306 0.459 –0.2552 –1.124 0.310 0.499 1.015 –0.417 –0.00187 0.000103 1.093 –0.0758 0.551
3.0 –2.2331 0.492 –0.2646 –1.121 0.310 0.499 1.014 –0.393 –0.00154 0.000084 1.090 –0.0737 0.562
4.0 –2.7975 0.507 –0.2738 –1.119 0.294 0.506 1.018 –0.386 –0.00135 0.000074 1.092 –0.0722 0.575
Source: Adapted from Campbell, K.W. in press. “Prediction of Strong Ground Motion Using the Hybrid Empirical Method: Example Application to Eastern North America,” Bull.
Seismol. Soc. Am.
5-45
0068_C05_fm Page 46 Thursday, August 1, 2002 7:59 AM
( )
ln Y = f 1 M S , r jb + f 2 ( S) (5.80)
( )
f 1 M S , r jb = c 1 + c 2 M S + c 3 ln R (5.81)
R = r jb2 + c 42 (5.82)
f 2 ( S) = c 5 SC + c 6 SD (5.83)
The regression coefficients and the standard deviation of lnY are listed in Table 5.17. The relation is
considered valid for MS = 4.0 to 7.5 and rjb ≤ 200 km. The relation predicts ground motion for building-
code Site Class B, equivalent to the condition SB = 1, unless SC or SD = 1, in which case it predicts ground
motion for Site Classes B or C, respectively (see Tables 5.2 and 5.4).
When an estimate of ground motion for a specified value of MW is required, it can be calculated from
the expression [Ambraseys et al., 1996]:
where seismic moment Mo is related to MW by Equation 5.6 and p is 0 for hhypo ≤ 30 km and 1 for hhypo
> 30 km. This equation, which was developed from European earthquakes, is very similar to one developed
by Ekstrom and Dziewonski [1988] from a database of worldwide earthquakes.
5.6.3.2 The Dahle et al. Model
The attenuation relation of Dahle et al. [1990] for the stable tectonic region of Europe can be
represented by the expression:
ln Y = c 1 + c 2 M S + ln R + c 3rhypo (5.85)
The regression coefficients and total standard deviation are listed in Table 5.18. The authors give no
guidelines for the use of the relation. However, the database contains recordings with MS = 3.0 to 6.9
and rhypo = 6 to 1300 km. The standard deviation of ln Y was separated into its intra- and inter-earthquake
components. The total standard deviation is reported here to be consistent with engineering convention.
One of the authors [Hilmar Bungum, personal communication, 2000] says that the relation can be used
to estimate ground motion for building-code Site Class B, equivalent to the condition SB = 1 (see
Tables 5.2 and 5.4).
The tectonic environment represented by the attenuation relation is described as an “intraplate” region,
another term for a stable tectonic region. The region is defined as an area that is tectonically stable and
geologically more uniform than a plate margin area. In order to obtain enough strong-motion recordings
for an analysis, the criterion for what qualified as intraplate was somewhat relaxed. Strong-motion
Source: Adapted from Ambraseys, N.N., Simpson, K.A., and Bommer, J.J. 1996. “Prediction of Horizontal Response
Spectra in Europe,” Earthquake Eng. Struct. Dyn., 25, 371–400.
recordings from ENA, Europe, China, and Australia were used, making this relation more worldwide
than strictly European. The largest amount of data comes from mainshock and aftershock recordings of
the 1976 Tangshan, China earthquake, the 1976 Friuli, Italy earthquake, and the 1988 Saguenay, Canada
earthquake. Some of the recordings do not belong to a genuine intraplate environment. This caveat
applies particularly to the Italian data and to a few recordings from southern Europe that come from an
area that has been characterized as a complex collision zone, where the inferred attenuation rate has
Source: Adapted from Dahle, A., Bungum, H., and Dvamme, L.G. 1990. “Atten-
uation Models Inferred from Intraplate Earthquake Recordings,” Earthquake Eng.
Struct. Dyn., 19, 1125–1141.
been shown to be somewhat stronger than that for a truly stable tectonic environment, such as north-
western Europe. The authors later applied a stricter criterion for selecting intraplate recordings that they
believe is more appropriate to the stable tectonic environment of northwestern Europe [Dahle et al.,
1991]. However, this latter relation incorporates only a limited number of periods for defining PSA and
is difficult to use in engineering practice.
The use of hypocentral distance as the distance measure implies that the attenuation relation is not strictly
valid for large earthquakes with extended rupture zones. In such cases, Boore [in press] suggests using rhypo
as an estimate of rrup . If an estimate of ground-motion in terms of MW is required, MS can be estimated
from MW using Equation 5.84 or a similar relation developed by Ekstrom and Dziewonski [1988].
5.6.4 Japan
Japan represents a unique tectonic environment where four large crustal plates converge to form an
archipelago (see Chapter 4). As a result of this unique tectonic environment, the strong-motion
recordings in Japan come from a mixture of shallow crustal earthquakes, subduction interface
earthquakes, and subduction intraslab or Wadati-Benioff earthquakes. Japanese scientists and engi-
neers do not generally distinguish between these different tectonic environments, other than to
include some measure of depth in their relations. An exception is the attenuation relations for PGA
and PGV developed by Si and Midorikawa [2000]. These authors, using Japanese recordings, and
Youngs et al. [1997] and Atkinson and Boore [2001], using worldwide recordings, have found that
the amplitudes and attenuation rates of ground motion vary significantly in these tectonic environ-
ments. As a result, some engineers also use attenuation relations derived from shallow crustal
earthquakes in WNA and from subduction earthquakes worldwide to estimate the ground motion
from similar tectonic environments in Japan. There are many attenuation relations that have been
developed specifically for Japan but very few of these predict PSA, because PGV is the principal
parameter used in the Japanese building codes. One exception is the attenuation relation developed
by Molas and Yamazaki [1995, 1996], which can be represented by the expression:
( )
ln Y = f 1 M J , rrup , hrup + f 2 ( S) (5.87)
( )
f 1 M J , rrup , hrup = c 1 + c 2 M J − ln rrup + c 3rrup + c 4hrup (5.88)
f 2 ( S) = c 5 S1 + c 6 S2 + c 7 S3 + c 8 S4 (5.89)
Source: Adapted from Molas, G.L. and Yamazaki, F. 1995. “Attenuation of Earthquake Ground Motion in Japan Including
Deep Focus Events,” Bull. Seismol. Soc. Am., 85, 1343–1358; and Molas, G.L. and Yamazaki, F. 1996. “Attenuation of Response
Spectra in Japan Using New JMA Records,” Bull. Earthquake Resistant Struct. Res. Center, 29, 115–128.
S1 Type 1 (rock) Tertiary or older rock (defined as bedrock) or diluvium with TS < 0.2
thickness < 10 m
S2 Type 2 (hard soil) Diluvium with thickness > 10 m or alluvium with thickness 0.2 ≤ TS < 0.4
< 10 m
S3 Type 3 (medium soil) Alluvium with thickness < 25 m, including soft layer with 0.4 ≤ TS < 0.6
thickness < 5 m
S4 Type 4 (soft soil) Other than above, usually soft alluvium or reclaimed land TS ≥ 0.6
Source: Adapted from Molas, G.L. and Yamazaki, F. 1995. “Attenuation of Earthquake Ground Motion in Japan Including
Deep Focus Events,” Bull. Seismol. Soc. Am., 85, 1343–1358; and Molas, G.L. and Yamazaki, F. 1996. “Attenuation of Response
Spectra in Japan Using New JMA Records,” Bull. Earthquake Resistant Struct. Res. Center, 29, 115–128.
The regression coefficients and the total standard deviation are listed in Table 5.19. The authors do
not give any guidelines for use of the relation. However, the database contains recordings with MJ = 4.0
to 7.8, rrup = 8 to 1000 km, and hrup ≤ 200 km. The standard deviation of lnY was separated into its intra-
and inter-earthquake components. Consistent with engineering convention, the total standard deviation
is reported here. The definition of the site parameters S1 through S4 is given in Table 5.20. Additional
guidance on their use is given in Table 5.4.
One limitation in the use of the relation is that the predicted strong-motion parameter goes to infinity
as rrup approaches zero. Because by definition rrup can be zero when the earthquake ruptures the Earth’s
surface, the authors admit that their attenuation relation is not valid for small rupture distances. This
limitation can be mitigated to some extent by replacing rrup by rhypo at short distances as recommended
by Boore [in press] when estimating ground motion using the stochastic method.
The relation combines earthquakes from shallow crustal earthquakes, subduction interface earth-
quakes, and subduction intraslab or Wadati-Benioff earthquakes. The only way to distinguish between
these different tectonic environments is through the depth parameter hrup. The functional form itself
does not allow for differences in amplitude between shallow crustal earthquakes and shallow subduction
interface earthquakes or in the rate of attenuation between shallow crustal earthquakes, subduction
interface, or subduction intraslab earthquakes. Youngs et al. [1997], Si and Midorikawa [2000], and
Atkinson and Boore [2001] have found these differences to be important.
Note: First value of PGA is for use as the short-period asymptote of PSA; second value of PGA is for use with
applications requiring PGA only [e.g., liquefaction analysis].
Source: Adapted from Crouse, C.B. 1991a. “Ground-Motion Attenuation Equations for Earthquakes on the Cascadia
Subduction Zone,” Earthquake Spectra, 7, 201–235; and Crouse, C.B. 1991b. “Erratum: Ground-Motion Attenuation
Equations for Earthquakes on the Cascadia Subduction Zone,” Earthquake Spectra, 7, 506.
ln Y = c 1 + c 2 M W + c 3 ln R + c 4hhypo (5.90)
R = rcer + c 5 exp (c 6 M W )
The regression coefficients and the standard deviation of ln Y are listed in Table 5.21. The author does
not give any guidance regarding use of the relation. However, the database contains recordings with MW
= 4.8 to 7.8, repi = 8 to 866 km, and hhypo = 0 to 238 km. There are two sets of regression coefficients for
PGA. According to the author [C.B. Crouse, personal communication, 2000], one set is based on the
same database that was used to derive the regression coefficients for PSA and should be used to calculate
the short-period asymptote of the pseudoacceleration response spectrum. The second set is based on a
larger database and should be used only when an estimate of PGA is needed.
The relation predicts ground motion for generic soil, equivalent to the condition SSoil = 1 (Table 5.4).
Because the recording sites are located throughout the world, they will not necessarily have an average
shear-wave velocity consistent with that of typical WNA generic soil sites. Nonetheless, this assumption
is usually made when using the relation to develop engineering estimates of strong ground motion in
the Pacific Northwest for earthquakes on the Cascadia subduction zone.
The distance measure rcer is defined as the distance to the center of energy release on the fault. For all
earthquakes less than magnitude 7.5, this point was assumed to be at the hypocenter of the earthquake.
For most of the larger events, the center of energy release was assumed to be at the centroid of the rupture
plane defined by the aftershocks. If special studies of source characteristics and aftershock distributions
were available, the center of energy release was selected as a point on the fault plane that corresponded
to the location of the greatest amount of energy release. This latter definition is not easily applied in
practice, because one does not generally know in advance where the maximum energy release will be.
Because rcer is essentially a point-source measure of distance, caution should be used when calculating
ground motion at short distances to the rupture plane of large earthquakes. This distance measure can
be close to zero or as large as a few hundred kilometers, depending on the size of the rupture plane, the
location of the site, and where on the fault plane the center of energy release is located. For most
engineering applications, rcer should be conservatively equated to the closest distance to the rupture
surface rrup, otherwise the ground motion predicted from the relation could be underestimated [C.B.
Crouse, personal communication, 2000].
The relation does not distinguish between subduction interface and subduction intraslab (Wadati-
Benioff) earthquakes. It does, however, include hypocentral depth as a parameter, which does account
to some extent for the higher ground motions that other investigators [Youngs et al., 1997; Si and
Midorikawa, 2000, Atkinson and Boore, 2001] have found for intraslab earthquakes.
5.6.5.2 The Youngs et al. Model
The attenuation relation of Youngs et al. [1997] can be represented by the expression:
( )
ln Y = f 1 M W , rrup , hhypo + f 2 ( zT ) (5.91)
( )
f 1 M W , rrup , hhypo = c 1 + c 2 M W + c 3 (10 − M W ) + c 4 ln R + c 5hhypo
2
(5.92)
R = rrup + c 7 exp (c 8 M W )
f 2 ( ZT ) = c 6 z T (5.93)
The parameter zT is 0 for subduction interface earthquakes and 1 for subduction intraslab or Wadati-
Benioff earthquakes.
The standard deviation of lnY was separated into its intra- and inter-earthquake components. Con-
sistent with engineering convention, only the total standard deviation is reported here. This total standard
deviation is given as a function of magnitude by the expression:
c + c M for M W ≤ 8.0
σ lnY = 9 10 W (5.94)
c11 for M W > 8.0
The regression coefficients are listed in Table 5.22. The relation is considered to be valid for MW ≥ 5.0
and rrup = 10 to 500 km. Although no depth criteria are given, the deepest earthquake in the database
has a hypocentral depth of 229 km. Differences in ground motion between generic soil, equivalent to
the condition SSoil = 1, and generic rock, equivalent to the condition SRock = 1 (Table 5.4), are taken into
Generic Rock
PGA 0.2418 1.414 0 –2.552 0.00617 0.3846 1.7818 0.554 1.45 –0.1 0.650
0.075 1.5168 1.414 0 –2.707 0.00617 0.3846 1.7818 0.554 1.45 –0.1 0.650
0.1 1.4298 1.414 –0.0011 –2.655 0.00617 0.3846 1.7818 0.554 1.45 –0.1 0.650
0.2 0.9638 1.414 –0.0027 –2.528 0.00617 0.3846 1.7818 0.554 1.45 –0.1 0.650
0.3 0.4878 1.414 –0.0036 –2.454 0.00617 0.3846 1.7818 0.554 1.45 –0.1 0.650
0.4 0.1268 1.414 –0.0043 –2.401 0.00617 0.3846 1.7818 0.554 1.45 –0.1 0.650
0.5 –0.1582 1.414 –0.0048 –2.360 0.00617 0.3846 1.7818 0.554 1.45 –0.1 0.650
0.75 –0.9072 1.414 –0.0057 –2.286 0.00617 0.3846 1.7818 0.554 1.45 –0.1 0.650
1.0 –1.4942 1.414 –0.0064 –2.234 0.00617 0.3846 1.7818 0.554 1.45 –0.1 0.650
1.5 –2.3922 1.414 –0.0073 –2.160 0.00617 0.3846 1.7818 0.554 1.50 –0.1 0.700
2.0 –3.0862 1.414 –0.0080 –2.107 0.00617 0.3846 1.7818 0.554 1.55 –0.1 0.750
3.0 –4.2692 1.414 –0.0089 –2.033 0.00617 0.3846 1.7818 0.554 1.65 –0.1 0.850
Generic Soil
PGA –0.6687 1.438 0 –2.329 0.00648 0.3648 1.097 0.617 1.45 –0.1 0.650
0.075 1.7313 1.438 –0.0019 –2.697 0.00648 0.3648 1.097 0.617 1.45 –0.1 0.650
0.1 1.8473 1.438 –0.0019 –2.697 0.00648 0.3648 1.097 0.617 1.45 –0.1 0.650
0.2 0.8803 1.438 –0.0019 –2.464 0.00648 0.3648 1.097 0.617 1.45 –0.1 0.650
Source: Adapted from Youngs, R.R., Chiou, S.J., Silva, W.J., and Humphrey, J.R. 1997. “Strong Ground Motion Attenuation Relationships
for Subduction Zone Earthquakes,” Seismol. Res. Lett., 68, 58–73.
account with different sets of regression coefficients. Because the recording sites are located throughout
the world, they will not necessarily have an average shear-wave velocity consistent with that of typical
WNA generic rock and soil sites. Nonetheless, this assumption is usually made when using the relation
to develop engineering estimates of strong ground motion in the Pacific Northwest for earthquakes on
the Cascadia subduction zone.
( )
ln Y = f 1 M W , r jb + f 2 ( S) (5.95)
( )
f 1 M W , r jb = c 1 + c 2 ( M W − 6) + c 3 ( M W − 6) + c 4 ln R
2
(5.96)
R = r jb2 + c 52 (5.97)
f 2 ( S) = c 6 SSoil (5.98)
The regression coefficients and the total standard deviation of ln Y are listed in Table 5.23. The relation
is considered to be valid for MW = 5.0 to 7.7 and rjb ≤ 100 km. The standard deviation of ln Y is separated
into its intra- and inter-earthquake components. The authors report the total standard deviation for the
horizontal component of ground motion both with and without including the randomness between the
two horizontal components. To be consistent with engineering convention, only the total standard
deviation excluding the component-to-component randomness is reported here.
The relation predicts ground motion for generic rock, equivalent to the condition SRock = 1. It is
evaluated for generic soil by setting SSoil = 1 (Table 5.4). Because the recordings are located throughout
the world, they will not necessarily have an average shear-wave velocity consistent with that of typical
WNA generic rock and soil sites. Nonetheless, this assumption is usually made when using this attenu-
ation relation to develop engineering estimates of strong ground motion in extensional regions in WNA.
Selected strong-motion recordings were not limited to extensional regions dominated by normal
faulting. In addition to recordings from eastern California, the Basin and Range province (Nevada, Idaho,
and Utah), Italy, and Greece, where large normal faults dominate the landscape, the authors also included
recordings from extensional regions with predominately strike-slip faulting. These latter regions include
the Imperial Valley of California, Baja California, Central America, New Zealand, and Turkey. Thus, there
is a large degree of crossover between this database and that used by other investigators who have
Source: Adapted from Spudich, P., Joyner, W.B., Lindh, A.G., Boore, D.M., Margaris, B.M., and Fletcher, J.B. 1999. “SEA99:
A Revised Ground Motion Prediction Relation for Use in Extensional Tectonic Regimes,” Bull. Seismol. Soc. Am., 89,
1156–1170.
developed attenuation relations that include shallow strike-slip earthquakes in many of these same
extensional environments.
One of the interesting consequences of including both normal-faulting and strike-slip earthquakes in
the relation is that the authors were able to evaluate possible differences in ground motion between these
two types of events. The results of their analysis indicated that there was no statistically significant difference
between ground motions from normal-faulting and strike-slip earthquakes in extensional regimes. This
conclusion is similar to assumptions made for non-extensional regimes by other investigators.
The authors compared their extensional recordings with predictions from the Boore et al. [1997]
attenuation relation evaluated for strike-slip faulting. They concluded that extensional rock ground
motion is on average 20% lower than the strike-slip ground motion predicted from the Boore et al.
model. Extensional soil recordings are generally less than 10% lower than ground motion predicted from
the Boore et al. model. An evaluation of the extensional relation by the authors found that it overpredicts
the extensional rock recordings, whereas it provides a reasonably unbiased estimate of the extensional
soil recordings. The authors concluded from this that their assumption of VS30 = 620 m/sec for the
extensional rock sites (Table 5.4) is probably too low.
Lower stress drops have been proposed by some investigators as a possible reason why ground motion
in an extensional environment should be lower than in a compressional environment [e.g., McGarr,
1984]. If this were the case for the Spudich et al. [1999] relation, the differences between the extensional
recordings and the Boore et al. [1997] empirical predictions should decrease rather than increase with
period as the authors have found [e.g., see Boore, in press]. This suggests an alternative hypothesis that
the differences between the Boore et al. and Spudich et al. ground-motion predictions may be due, at
least in part, to systematic differences in crustal properties or local site conditions between WNA and
non-WNA recordings rather than to lower stress drops.
[ ( ) ( )
ln YHW , FW = ln Y + f 1 ( M W ) f 2 rrup FW + f 3 rrup HW ] (5.99)
where Y is the average horizontal or vertical component of PGA or PSA excluding the hanging-wall
or footwall effects, YHF,FW is the value of Y including these effects, and
0 for M W ≤ 5.5
f 1 ( M W ) = M W − 5.5 for 5.5 < M W < 6.5 (5.100)
1 for M W ≥ 6.5
0 for rrup ≤ 6 km
r −6
c rup for 6 < rrup ≤ 12 km
1 6
( )
f 2 rrup
= c 1
rrup − 25
for 12 < rrup ≤ 25 km (5.101)
0 for rrup ≤ 4 km
c 2
(
π r −4
cos rup )
+ π + 1 for 4 < rrup ≤ 8 km
2 4
( )
f 3 rrup = c 2
for 8 < rrup ≤ 18 km (5.102)
c 2 (
π r − 18
cos rup )
+ 1 for 18 < rrup ≤ 25 km
2 7
0 for rrup > 25 km
The regression coefficients are listed in Table 5.24. The definition of the hanging wall and footwall to
be used with the model is schematically shown in Figure 5.10. Somerville and Abrahamson [1995, 2000]
suggest that the hanging-wall and footwall effects are primarily geometrical effects that result from the
use of rrup as the distance measure. A similar geometrical effect is expected from the distance measure
rseis used by Campbell [1997] and Campbell and Bozorgnia [in press]. According to Somerville and
Abrahamson, the rjb distance measure used by Boore et al. [1997] implicitly accounts for the difference
in ground motion over the hanging wall and footwall and, in their opinion, should not be corrected for
these geometrical effects. The hanging-wall and footwall effects given by Equations 5.99 to 5.102 and by
22.5°
FIGURE 5.10 Definition of hanging wall and footwall in the engineering model used to incorporate hanging-wall
and footwall effects in the estimation of strong ground motion. (From Somerville, P. and Abrahamson, N. 1995.
“Ground Motion Prediction for Thrust Earthquakes,” in Proc. SMIP95 Seminar on Seismological and Engineering
Implications of Recent Strong-Motion Data, May 16, San Francisco, pp. 11–23. California Strong Motion Instrumen-
tation Program, Sacramento.)
Abrahamson and Silva [1997] can theoretically be applied to any attenuation relation that uses rrup or
rseis as the distance measure. Because they are geometrical effects, they should theoretically be valid for
both reverse-faulting and normal-faulting earthquakes. However, there is insufficient data to test this
hypothesis for normal-faulting events.
The model predicts that sites located on the hanging wall at distances of 8 to 18 km can have an
increase in ground motion of up to 45% for the horizontal component of ground motion and 88% for
the vertical component of ground motion. Sites on the footwall at distances of 12 to 25 km can have a
decrease in ground motion of up to 47%. Although these results are largely driven by recordings from
the Northridge earthquake, similar effects were observed during the 1999 MW = 7.6 Chi-Chi, Taiwan
earthquake [Loh, 1999; Tsai and Huang, 2000], where some sites on the hanging wall had PGA values
ranging from 0.5 to 1 g and PGV values ranging from 100 to 300 cm/sec. The largest values of PGV were
recorded near the surface trace of the fault, where they may also have been impacted by source directivity.
Abrahamson and Silva [1997] and Somerville and Abrahamson [2000] recommend against incorporating
footwall effects in the engineering estimation of strong ground motion because, in their opinion, these
effects are not as credible as the hanging-wall effects. However, the apparent footwall effects observed
during the 1999 Chi-Chi earthquake, if proven to be valid, will add to their credibility.
( ) (
ln YDir = ln Y + f 1 ( DR, ξ) T rrup T ( M W ) + f 2 rrup , M W , ξ ) (5.103)
where Y is the average horizontal component of PGA or PSA excluding directivity effects, YDir is the
value of Y including these effects, and for strike-slip faulting:
and
1
[ ( )
2 (cos 2ξ) c 3 + c 4 ln rrup + 1 + c 5 ( M W − 6)
] for fault - normal
( )
1
2
[ ( )
f 2 rrup , M W , ξ = − (cos 2ξ) c 3 + c 4 ln rrup + 1 + c 5 ( M W − 6) ] for fault - parallel (5.106)
0 for ξ ≥ 45°
1 for rrup ≤ 30 km
( )
T rrup
0
(
= 1 − rrup − 30 30 ) for 30 < rrup < 60 km (5.107)
for rrup ≥ 60 km
1 for M W ≥ 6.5
T ( M W ) = 1 − (6.5 − M W ) 0.5 for 6.0 < M W < 6.5 (5.108)
0 for M W ≤ 6.0
The standard deviation of the predicted strong-motion parameter when directivity effects are taken
into account can be calculated from the expression:
where σlnY, Dir is the standard deviation of ln YDir, and σln Y is the standard deviation of lnY.
In the above equations, the length and width ratio DR = s/L and d/W are defined as the fraction of
fault rupture length L and fault rupture width W that ruptures towards the site for strike-slip faults and
dip-slip faults, respectively; and ξ = θ and φ are the azimuth or zenith angle between the fault rupture
plane and the ray path to the site for strike-slip and dip-slip faults, respectively. These parameters are
defined schematically in Figure 5.11. The regression coefficients are listed in Table 5.25. Note that in this
table the values of c1 and c2 depend on the faulting mechanism, where dip slip is a generic term for both
reverse, thrust, and normal faulting.
FIGURE 5.11 Definition of fault-rupture directivity parameters used in the engineering model used to incorporate
rupture directivity effects in the estimation of strong ground motion. (From Somerville, P.G., Smith, N.F., Graves,
R.W., and Abrahamson, N.A. 1997. “Modification of Empirical Strong Ground Motion Attenuation Relations to
Include the Amplitude and Duration Effects of Rupture Directivity,” Seismol. Res. Lett., 68, 199–222. With permission.)
TABLE 5.25 Coefficients for the Engineering Model of Rupture Directivity Effects
Strike Slip Dip Slip
Tn (s) c1 c2 c1 c2 c3 c4 c5
Source: Adapted from Somerville, P.G., Smith, N.F., Graves, R.W., and Abrahamson, N.A. 1997.
“Modification of Empirical Strong Ground Motion Attenuation Relations to Include the Amplitude
and Duration Effects of Rupture Directivity,” Seismol. Res. Lett., 68, 199–222; and Abrahamson, N.A.
2000. “Effects of Rupture Directivity on Probabilistic Seismic Hazard Analysis,” in Proc. 6th Inter-
national Conference on Seismic Zonation, Nov. 12–15, Palm Springs, CA, Proc. CD-ROM, Earthquake
Engineering Research Institute, Oakland, CA.
According to the relation, maximum directivity conditions for strike-slip faulting occur when (s/L)
cosθ = 1 and can result in an increase of up to 68% in the average horizontal component of PSA at a
period of 5 sec. Minimum directivity conditions occur when (s/L) cosθ = 0 and can result in a similar
reduction in PSA. These effects are only about ±36% for dip-slip faulting. Maximum fault-normal and
fault-parallel effects occur at large magnitudes and long periods when θ = φ = 0. These effects can result
in an increase of up to 107% in the fault-normal component of PSA and a similar decrease in the fault-
parallel component. Because the spatial and fault-normal effects are cumulative, the total maximum
directivity effects can approach a factor of 3.5.
TABLE 5.26 Seismological Parameters Used in the Evaluation of the Engineering Models
Faulting Scenario F FRV FTH HW FW ZT rjb rrup rseis rhypo rcer hhypo
TABLE 5.27 Site Parameters Used in the Evaluation of the Engineering Models
Attenuation Relation Generic Rock Generic Soil Firm/Hard Rock
Note: For those attenuation relations that represent only one site condition or whose reference site
condition is not represented by a specific site parameter, the name of the reference site class or the
approximate value of the 30-m velocity is given in parentheses. Blank cells indicate that the attenuation
relation was not used to evaluate that site condition.
faulting scenario is kept simple to avoid unnecessary complication, although it should be noted that
actual scenarios can be much more difficult to evaluate. The definition of the faulting scenarios and the
seismological parameters that were used to evaluate the different engineering models are listed in
Table 5.26. The values of the site parameters that were used to evaluate each of these models are listed
in Table 5.27.
Predictions of response spectra for shallow active tectonic regions in WNA, Japan, and worldwide
extensional regions are done for generic rock (nominally VS30 = 620 m/sec). Most engineering estimates
of ground motion are done for this site condition so that soil effects can be treated in a site-specific
manner using dynamic site-response analyses or building-code site factors. No adjustment is made to
account for possible variations in VS30 between different definitions of generic rock, which is also typical
of engineering evaluations. Predictions for shallow stable tectonic regions in ENA are done for ENA hard
rock (nominally VS30 = 2800 m/sec). Those in Europe are done for building-code Site Class B, the reference
site condition for the Dahle et al. [1990] attenuation relation. Site Class B has a maximum 30-m velocity
of only 1500 m/sec (Table 5.2) but it is the closest site condition to ENA hard rock that is available for
evaluating the attenuation relation for the European stable tectonic region. Where predicted spectra for
nominal hard rock in WNA are compared to those for ENA hard rock, WNA firm rock (nominally
VS30 = 760 m/sec) is used for this comparison. This latter value for the 30-m velocity is used by Frankel
1.0 1.0
0.1 0.1
0.0 0.0
0.01 0.1 1 10 0.01 0.1 1 10
Period (sec) Period (sec)
(c) (d)
Horizontal Spectral Acceleration (g)
Horizontal Spectral Acceleration (g)
1.0
1.0
0.1
0.1
0.0 0.0
0.01 0.1 1 10 0.01 0.1 1 10
Period (sec) Period (sec)
FIGURE 5.12 Predicted horizontal acceleration response spectra for shallow active and stable tectonic regions of
North America for MW = 7.0 and a nominal surface distance (rjb) of 10 km (0 km over the hanging wall) showing
differences among: (a) the four selected attenuation relations for generic rock in WNA; (b) the effects of faulting
mechanism and fault geometry for generic rock in WNA; (c) the effects of rupture directivity for generic rock in
WNA; and (d) the three selected attenuation relations for very hard rock in ENA (see Tables 5.26 and 5.27 for a
description of the seismological and site parameters used in the evaluations).
et al. [1996, 2000] to develop the seismic hazard maps adopted for use with the 1997 NEHRP and 2000
International Building Codes [BSSC, 1998; ICC, 2000]. Because the Crouse [1991a, 1991b] relation is
valid only for generic soil, predicted spectra for the subduction zone environments both worldwide and
in Japan are calculated for this site condition (nominally VS30 = 310 m/sec).
Figure 5.12 compares the predicted response spectra for North American active (WNA) and stable
(ENA) tectonic regions. Except for those plots where individual attenuation relations are identified, the
spectra for WNA and ENA represent the geometric mean of the individual estimates from the relations
for each of these regions. The WNA comparisons also show differences due to faulting mechanism,
hanging-wall, footwall, and source directivity effects. Directivity effects are evaluated for (s/L) cosθ = 1
and for the fault-normal (FN) and fault-parallel (FP) horizontal component. Unless otherwise noted,
all evaluations are for an unknown or random faulting mechanism and for average directivity effects.
The large difference in the spectral accelerations at long periods in the ENA comparison clearly shows
the large difference between the one-corner source spectrum used by Toro and coworkers [1997] and
the two-corner source spectrum used by Atkinson and Boore [1995, 1997].
Figure 5.13 compares the predicted response spectra for worldwide active and stable tectonic regions
and for worldwide subduction interface and intraslab (Wadati-Benioff) earthquakes. Comparisons for
active tectonic regions include WNA, Europe, Japan, and worldwide extensional environments. Note that
the predicted response spectrum for the extensional regime is not that much less than the predicted
response spectra for the other regions. This difference would even be less if an adjustment was made for
the bias noted by Spudich and coworkers[1999] for their extensional rock predictions. Comparisons for
stable tectonic regions include ENA and Europe, with WNA hard rock included to show the significant
difference in spectral shape between ENA and WNA. The large differences in these latter spectra are due
to two factors: (1) differences in the tectonic regime in which the European stable tectonic region is
intermediate to the ENA stable tectonic and WNA active tectonic regions [Dahle et al., 1990], and
(2) differences in local site conditions in which 30-m velocity is approximately 2800 m/sec for ENA,
1130 m/sec for Europe, and 760 m/sec for WNA. Comparisons for subduction zones include Japan and
worldwide interface and intraslab events. Note that in the latter comparison, the Crouse [1991a, 1991b]
and Molas and Yamazaki [1995, 1996] attenuation relations include hypocentral depth as the only param-
eter to distinguish between interface and intraslab events, which leads them to predict smaller intraslab
estimates than the Youngs et al. [1997] relation. The large variability in the subduction zone spectra
demonstrates the large degree of uncertainty associated with the estimation of ground motion in sub-
duction zone environments.
There is one last item — vertical response spectra — that should be mentioned. Vertical ground
motion has largely been ignored by engineers because current practice dictates that (1) it is usually so
small as to not add significantly to the combined vertical and horizontal seismic design loads, (2) the
structure is already designed for a vertical load of 1 g and should not need strengthening in that direction,
and (3) its peak is usually out of phase with the peak in the horizontal ground motion. This thinking
came from an often-quoted engineering “rule of thumb” that vertical motion can be assumed to be two
thirds that of the horizontal motion [e.g., Newmark and Hall, 1982]. None of these conditions is
appropriate at near-source distances from large earthquakes in many cases. The recent plethora of near-
source ground motion recordings has shown that vertical response spectra can have short-period
amplitudes that exceed the average horizontal amplitudes by a factor of two or more, especially for sites
on soil, and that this difference strongly increases with decreasing distance [Campbell and Bozorgnia,
in press]. Figure 5.14 compares the vertical and average horizontal response spectra predicted from
those WNA attenuation relations that provide estimates for both [Abrahamson and Silva, 1997; Camp-
bell, 1997, 2000c, 2001; Sadigh et al., 1993]. The comparison is made using the same magnitude and
distance as that used in the comparison of the horizontal spectra. These spectra represent generic rock,
for which the vertical-to-horizontal (V/H) spectral ratio is minimal. Figure 5.15 shows the large differ-
ence in V/H spectral ratio that can be expected for soil (approximately building-code Site Class D) for
distances ranging from 3 to 60 km. This figure clearly shows that the largest V/H ratios occur at periods
of 0.05 to 0.15 sec but that this ratio decreases rapidly with period, reaching values less than one half
at periods exceeding 0.3 sec. Therefore, the two thirds “rule of thumb” is not appropriate for longer
periods either, where it is much less than commonly assumed.
0.1
0.1
0.0 0.0
0.01 0.1 1 10 0.01 0.1 1 10
Period (sec) Period (sec)
1.0 1.0
0.1 0.1
0.0 0.0
0.01 0.1 1 10 0.01 0.1 1 10
Period (sec) Period (sec)
FIGURE 5.13 Predicted horizontal acceleration response spectra for MW = 7.0 and a nominal surface distance (rjb)
of 10 km showing differences among: (a) four worldwide shallow active tectonic regions for generic rock or its
equivalent; (b) two worldwide stable tectonic regions together with the active tectonic region of WNA (for compar-
ison) for the hardest rock available in each region; (c) the three selected subduction interface relations for generic
soil or its equivalent; and (d) the three selected subduction intraslab (Wadati-Benioff zone) relations for generic soil
or its equivalent (see Tables 5.26 and 5.27 for a description of the seismological and site parameters used in the
evaluations).
1.0
Average Horizontal
Vertical
0.0
0.01 0.1 1 10
Period (sec)
FIGURE 5.14 Predicted horizontal and vertical acceleration response spectra for shallow active tectonic regions in
WNA for MW = 7.0 and a nominal surface distance (rjb) of 10 km (see Tables 5.26 and 5.27 for a description of the
seismological and site parameters used in the evaluations).
1.8
STRIKE SLIP, M 6.5
1.6 R = 3 km
R = 10 km
R = 20 km
1.4 R = 60 km
R = 3 km
R = 10 km
V/H Ratio
1.2 R = 20 km
R = 60 km
1.0
0.8
0.6
0.4
0.2
0.01 0.1 1 10
Period (sec)
FIGURE 5.15 The effect of distance on the vertical-to-horizontal response spectral ratio for recordings on firm
soil. Symbols represent the ratio for PGA. (From Campbell, K.W. and Bozorgnia, Y. 1999. “Vertical Ground Motion:
Characteristics, Relationship with Horizontal Component, and Building-Code Implications,” in Proc. SMIP99 Sem-
inar on Utilization of Strong-Motion Data, M. Huang, Ed., Sept. 15, San Francisco, pp. 23–49. California Strong
Motion Instrumentation Program, Sacramento, CA).
Defining Terms
Aleatory variability — Uncertainty due to randomness in the phenomena, as differentiated from
epistemic uncertainty.
Anelastic attenuation — The diminution of ground motion with distance from the source due to
material damping and scattering of waves from inhomogeneities in the crust.
Attenuation relation — An equation or tabulation used to estimate a strong-motion parameter from
one or more seismological parameters; also known as a ground motion relation.
Basement rock — The more resistant, generally crystalline rock that lies beneath layers or irregular
deposits of younger, relatively deformed sedimentary rock.
Critical reflection — The incidence angle below which the ground-motion ray is completely
reflected off a layer of higher wave velocity.
Epicenter — The point on the Earth’s surface directly above the hypocenter.
Epistemic uncertainty — Uncertainty due to lack of knowledge (e.g., omission of a key parameter
due to not recognizing its importance in a phenomenon, uncertainty in a parameter or
model).
Faulting mechanism — The type or style of faulting defined by the direction of slip on the fault
rupture plane; usually referred to by such terms as strike slip, reverse, thrust, normal, or
oblique.
Focus — See hypocenter.
Footwall — That portion of the crust that lies below the fault or fault rupture plane.
Fourier spectrum — A complete frequency-domain description of the ground motion time history
a(t) given by
S
F (ω ) =
∫ a(t )exp (−iωt )dt ,
0
1/ 2
S
2
S
2
F (ω ) =
∫
0
a(t ) cos ωt dt +
∫
0
a (t ) sin ωt dt
.
Frequency — The reciprocal of period, i.e., the number of cycles of oscillation per unit of time (e.g.,
1 sec). Usually measured in terms of hertz (1 Hz = 1 cycle per second).
Geometric attenuation — The diminution of ground motion with distance from the source as the
area of the wave front expands.
Ground motion — The vibration of the ground in the time or frequency domain measured by a
seismometer that records acceleration, velocity, or displacement, or an estimate of this
vibration or a ground-motion parameter that characterizes this vibration.
Hanging wall — That portion of the crust that lies above the fault or fault rupture plane.
Hypocenter — The point within the Earth where the earthquake rupture begins (see also focus).
Local site conditions — A qualitative or quantitative description of the material properties of the
soil and sedimentary rock layers above basement rock.
Magnitude — An instrumental or seismological measure of an earthquake’s size proportional to the
logarithm of the amplitude or energy of ground motion.
Natural frequency — The reciprocal of natural period.
Natural period — The period of an oscillator or structure during free (i.e., unforced) vibration.
Period — The duration of time (e.g., number of seconds) required to complete one oscillation.
Radiation pattern — A geometric description of the amplitude of ground motion and the sense of
initial motion at the source which for shear waves has a low-order symmetry that can be
used to infer the faulting mechanism.
Rake angle — The angle between the direction of slip on the fault rupture plane and the fault strike.
Response spectrum — A plot of undamped natural period or frequency vs. the maximum response
of a viscously damped single-degree-of-freedom (SDOF) oscillator subjected to a specified
ground motion time history at its base.
Seismogenic — That part of the Earth’s crust that is capable of generating ground motion at periods
of engineering interest, usually 10 sec or less.
Seismometer — An instrument used to record ground motion.
Seismological parameter — A parameter used to characterize a seismological property of the earth-
quake source, the propagation medium, or the response of the materials beneath the site.
Shear-wave velocity — The speed at which shear waves travel through a material; shear waves are
waves whose amplitude is perpendicular to the direction of propagation and are the most
potentially damaging to man-made structures.
Source directivity — The azimuthal perturbation of the radiation pattern due to rupture propaga-
tion on the fault in which the amplitude increases in the direction of rupture and decreases
in the opposite direction.
Stress drop — The amount of stress released at the rupture front during an earthquake.
Strike — The orientation of a fault on the Earth’s surface, usually measured clockwise from north.
Strong ground motion — Ground motion having the potential to cause measurable damage to a
structure’s architectural or structural components; usually associated with a PGA of 0.05 g
or greater.
Strong-motion parameter — A parameter characterizing the amplitude of strong ground motion
in the time domain (time-domain parameter) or the frequency domain (frequency-domain
parameter).
Time history — A data set, usually composed of one vertical and two orthogonal horizontal com-
ponents, describing a strong-motion parameter (such as ground acceleration) as a function
of time.
Tectonic environment — The type of tectonic deformation that occurs in a region, usually described
by such terms as active, stable, compressional, extensional, or subduction.
References
Abrahamson, N.A. 2000. “Effects of Rupture Directivity on Probabilistic Seismic Hazard Analysis,”
in Proc. 6th International Conference on Seismic Zonation, Nov. 12–15, Palm Springs, CA, Proc.
CD-ROM, Earthquake Engineering Research Institute, Oakland.
Abrahamson, N.A. and Shedlock, K.M. 1997. “Overview,” Seismol. Res. Lett., 68, 9–23.
Abrahamson, N.A. and Silva, W.J. 1997. “Empirical Response Spectral Attenuation Relations for
Shallow Crustal Earthquakes,” Seismol. Res. Lett., 68, 94–127.
Abrahamson, N.A. and Somerville, P.C. 1996. “Effects of the Hanging Wall and Footwall on Ground
Motions Recorded during the Northridge Earthquake,” Bull. Seismol. Soc. Am., 86, S93–S99.
Abrahamson, N.A. and Youngs, R.R. 1992. “A Stable Algorithm for Regression Analyses Using the
Random Effects Model,” Bull. Seismol. Soc. Am., 82, 505–510.
Ambraseys, N.N., Simpson, K.A., and Bommer, J.J. 1996. “Prediction of Horizontal Response Spectra
in Europe,” Earthquake Eng. Struct. Dyn., 25, 371–400.
Anderson, J.G. 2000. “Expected Shape of Regressions for Ground Motion Parameters on Rock,” Bull.
Seismol. Soc. Am., 90, S43–S52.
Anderson, J.G., in press. “Strong Motion Seismology,” in International Handbook of Earthquake and
Engineering Seismology, W.H.K. Lee, H. Kanamori, P.C. Jennings, and C. Kisslinger, Eds., Chapter
57, Academic Press, London.
Anderson, J.G. and Hough, S.E. 1984. “A Model for the Shape of the Fourier Amplitude Spectrum
of Acceleration at high Frequencies,” Bull. Seismol. Soc. Am., 74, 1969–1993.
Ansary, M.A., Yamazaki, F., and Katayama, T. 1995. “Statistical Analysis of Peaks and Directivity of
Earthquake Ground Motion,” Earthquake Eng. Struct. Dyn., 24, 1527–1539.
Atkinson, G.M. 1995. “Optimal Choice of Magnitude Scales for Seismic Hazard Analysis,” Seismol.
Res. Lett., 66, 51–55.
Atkinson, G.M. and Beresnev, I. 1997. “Don’t Call it Stress Drop,” Seismol. Res. Lett., 68, 3–4.
Atkinson, G.M. and Boore, D.M. 1995. “New Ground Motion Relations for Eastern North America,”
Bull. Seismol. Soc. Am., 85, 17–30.
Atkinson, G.M. and Boore, D.M. 1997. “Some Comparisons between Recent Ground Motion Rela-
tions,” Seismol. Res. Lett., 68, 24–40.
Atkinson, G.M. and Boore, D.M. 1998. “Evaluation of Models for Earthquake Source Spectra in
Eastern North America,” Bull. Seismol. Soc. Am., 88, 917–934.
Atkinson, G.M. and Boore, D.M. 2001. “Empirical Ground Motion Relations for Subduction Zone
Earthquakes (Abstract),” Seismol. Res. Lett., 72, 271.
Beresnev, I.A. and Atkinson, G.M. 1998. “FINSIM — a FORTRAN Program for Simulating Stochastic
Acceleration Time Histories from Finite Faults,” Seismol. Res. Lett., 69, 27–32.
Boatwright, J.A. and Boore, D.M. 1982. “Analysis of the Ground Accelerations Radiated by the 1980
Livermore Valley Earthquakes for Directivity and Dynamic Source Characteristics,” Bull. Seis-
mol. Soc. Am., 72, 1843–1865.
Bolt, B.A. 1993. Earthquakes and Geological Discovery, Scientific American Library, New York.
Boore, D.M 1996. “SMSIM — Fortran Programs for Simulating Ground Motions from Earthquakes:
Version 1.0,” U.S. Geological Survey, Open-File Rept. 96-80-A and 96-80-B.
Boore, D.M. 2001. “Comparisons of Ground Motions from the 1999 Chi-Chi Earthquake with
Empirical Predictions Largely Based on Data from California,” Bull. Seismol. Soc. Am., 91,
1212–1217.
Boore, D.M., in press. “Prediction of Ground Motion Using the Stochastic Method,” Pure Appl. Geophys.
Boore, D.M. and Atkinson, G.M. 1987. “Stochastic Prediction of Ground Motion and Spectral
Response Parameters at Hard-Rock Sites in Eastern North America,” Bull. Seismol. Soc. Am.,
77, 440–467.
Boore, D.M. and Joyner, W.B. 1984. “A Note on the Use of Random Vibration Theory to Predict
Peak Amplitudes of Transient Signals,” Bull. Seismol. Soc. Am., 74, 2035–2039.
Boore, D.M. and Joyner, W.B. 1991. “Estimation of Ground Motion at Deep-Soil Sites in Eastern
North America,” Bull. Seismol. Soc. Am., 81, 2167–2185.
Boore, D.M. and Joyner, W.B. 1997. “Site Amplification for Generic Rock Sites,” Bull. Seismol. Soc.
Am., 87, 327–341.
Boore, D.M., Joyner, W.B., and Fumal, T.E. 1993. “Estimation of Response Spectra and Peak Accel-
erations from Western North American Earthquakes: An Interim Report,” U.S. Geological
Survey, Open-File Rept. 93-509.
Boore, D.M., Joyner, W.B., and Fumal, T.E. 1994. “Estimation of Response Spectra and Peak Accel-
erations from Western North American Earthquakes: An Interim Report, Part 2,” U.S. Geo-
logical Survey, Open-File Rept. 94-127.
Boore, D.M., Joyner, W.B., and Fumal, T.E. 1997. “Equations for Estimating Horizontal Response
Spectra and Peak Acceleration from Western North American Earthquakes: A Summary of
Recent Work,” Seismol. Res. Lett., 68, 128–153.
Boore, D.M., Joyner, W.B., and Wennerberg, L. 1992. “Fitting the Stochastic ω−2 Source Model to
Observed Response Spectra in Western North America: Trade-Offs between ∆σ and κ,” Bull.
Seismol. Soc. Am., 82, 1956–1963.
Borcherdt, R.D. 1994. “Estimates of Site-Dependent Response Spectra for Design (Methodology and
Justification),” Earthquake Spectra, 10, 617–653.
Brillinger, D.R. and Preisler, H.K. 1984. “An Exploratory Analysis of the Joyner-Boore Attenuation
Data,” Bull. Seismol. Soc. Am., 74, 1441–1450.
Brune, J. 1970. “Tectonic Stress and the Spectra of Seismic Shear Waves,” J. Geophys. Res., 75,
4997–5009.
Brune, J. 1971. “Correction: Tectonic Stress and the Spectra of Seismic Shear Waves,” J. Geophys.
Res., 76, 5002.
Brune, J.N. 1999. “Precarious Rocks Along the Mojave Section of the San Andreas Fault, California:
Constraints on Ground Motion from Great Earthquakes,” Seismol. Res. Lett., 70, 29–33.
Brune, J.N. 2001. “Shattered Rock and Precarious Rock Evidence for Strong Asymmetry in Ground
Motions during Thrust Faulting,” Bull. Seismol. Soc. Am., 91, 441–447.
BSSC (Building Seismic Safety Council). 1998. NEHRP Recommended Provisions for Seismic Regulations
for New Buildings and Other Structures, 1997 ed., Rept. FEMA-302, Federal Emergency Manage-
ment Agency, Washington, D.C.
Campbell, K.W. 1981. “Near-Source Attenuation of Peak Horizontal Acceleration,” Bull. Seismol. Soc.
Am., 71, 2039–2070.
Campbell, K.W. 1985. “Strong Motion Attenuation Relations: A Ten-Year Perspective,” Earthquake
Spectra, 1, 759–804.
Campbell, K.W. 1986. “Empirical Prediction of Free-Field Ground Motion Using Statistical Regres-
sion Models,” in Proc. Soil Structure Interaction Workshop, NUREG/CP-0054, U.S. Nuclear Reg-
ulatory Commission, Washington, D.C., pp. 72–115.
Campbell, K.W. 1987. “Predicting Strong Ground Motion in Utah,” in Assessment of Regional Earth-
quake Hazards and Risk Along the Wasatch Front, Utah, Vol. II, P.L. Gori and W.W. Hays, Eds.,
U.S. Geological Survey, Open-File Rept. 87-585, pp. L1–L90.
Campbell, K.W. 1997. “Empirical Near-Source Attenuation Relationships for Horizontal and Vertical
Components of Peak Ground Acceleration, Peak Ground Velocity, and Pseudo-Absolute
Acceleration Response Spectra,” Seismol. Res. Lett., 68, 154–179.
Campbell, K.W. 1998. “Empirical Analysis of Peak Horizontal Acceleration, Peak Horizontal Velocity,
and Modified Mercalli Intensity,” in The Loma Prieta, California, Earthquake of October 17, 1989
— Earth Structures and Engineering Characterization of Ground Motion, T.L. Holzer, Ed., U.S.
Geological Survey, Professional paper 1552-D, pp. D47–D68.
Campbell, K.W. 2000a. “Seismology, Engineering,” in Encyclopedia of Physical Science and Technology,
3rd ed., R.A. Meyers, Ed., Academic Press, San Diego.
Campbell, K.W. 2000b. “Predicting Strong Ground Motion in Utah,” in Assessment of Regional Earth-
quake Hazards and Risk Along the Wasatch Front, Utah, P.L. Gori and W.W. Hays, Eds., U.S.
Geological Survey, Professional Paper 1500-L, pp. L1–L31.
Campbell, K.W. 2000c. “Erratum: Empirical Near-Source Attenuation Relationships for Horizontal
and Vertical Components of Peak Ground Acceleration, Peak Ground Velocity, and Pseudo-
Absolute Acceleration Response Spectra,” Seismol. Res. Lett., 71, 353–355.
Campbell, K.W. 2001. “Erratum: Empirical Near-Source Attenuation Relationships for Horizontal
and Vertical Components of Peak Ground Acceleration, Peak Ground Velocity, and Pseudo-
Absolute Acceleration Response Spectra,” Seismol. Res. Lett., 72, 474.
Campbell, K.W., in press. “Prediction of Strong Ground Motion Using the Hybrid Empirical Method:
Example Application to Eastern North America,” Bull. Seismol. Soc. Am.
Campbell, K.W. and Bozorgnia, Y. 1994. “Empirical Analysis of Strong Ground Motion from the
1992 Landers, California, Earthquake,” Bull. Seismol. Soc. Am., 84, 573–588.
Campbell, K.W. and Bozorgnia, Y. 1999. “Vertical Ground Motion: Characteristics, Relationship
with Horizontal Component, and Building-Code Implications,” in Proc. SMIP99 Seminar on
Utilization of Strong-Motion Data, M. Huang, Ed., Sept. 15, San Francisco, pp. 23–49, California
Strong Motion Instrumentation Program, Sacramento.
Campbell, K.W. and Bozorgnia, Y. 2000. “New Empirical Models for Predicting Near-Source Hori-
zontal, Vertical, and V/H Response Spectra: Implications for Design,” in Proc. 6th International
Conference on Seismic Zonation, Nov. 12–15, Palm Springs, CA, Proc. CD-ROM, Earthquake
Engineering Research Institute, Oakland, CA.
Campbell, K.W. and Bozorgnia, Y., in press. “Updated Near-Source Ground Motion (Attenuation) Rela-
tions for the Horizontal and Vertical Components of Peak Ground Acceleration and Acceleration
Response Spectra,” Bull. Seismol. Soc. Am.
Cartwright, D.E. and Longuet-Higgins, M.S. 1956. “The Statistical Distribution of the Maxima of a
Random Function,” Proc. R. Soc. London, Ser. A, 237, 212–232.
Crouse, C.B. 1991a. “Ground-Motion Attenuation Equations for Earthquakes on the Cascadia Sub-
duction Zone,” Earthquake Spectra, 7, 201–235.
Crouse, C.B. 1991b. “Erratum: Ground-Motion Attenuation Equations for Earthquakes on the Cas-
cadia Subduction Zone,” Earthquake Spectra, 7, 506.
Dahle, A., Bungum, H., and Dvamme, L.G. 1990. “Attenuation Models Inferred from Intraplate
Earthquake Recordings,” Earthquake Eng. Struct. Dyn., 19, 1125–1141.
Dahle, A., Bungum, H., and Dvamme, L.G. 1991. “Empirically Derived PSV Spectral Attenuation
Models for Intraplate Conditions,” European Earthquake Eng., 3, 42–52.
Donovan, N.C. and Bornstein, A.E. 1978. “Uncertainties in Seismic Risk Procedures,” J. Geotech. Eng.
Div. ASCE, 104, 869–887.
Draper, N.R. and Smith, H. 1981. Applied Regression Analysis, 2nd ed., John Wiley & Sons, New York.
Ekstrom, G. and Dziewonski, A.M. 1988. “Evidence of Bias in Estimation of Earthquake Size,” Nature,
332, 319–322.
EPRI (Electric Power Research Institute). 1993. “Methods and Guidelines for Estimating Earthquake
Ground Motion in Eastern North America,” in Guidelines for Determining Design Basis Ground
Motions, Vol. 1, Rept. EPRI TR-102293, Electric Power Research Institute, Palo Alto, CA.
Field, E.H. 2000. “A Modified Ground-Motion Attenuation Relationship for Southern California
that Accounts for Detailed Site Classification and a Basin-Depth Effect,” Bull. Seismol. Soc. Am.,
90, S209–S221.
Field, E.H. and the SCEC Phase III Working Group. 2000. “Accounting for Site Effects in Probabilistic
Seismic Hazard Analysis of Southern California: Overview of the SCEC Phase III Report,” Bull.
Seismol. Soc. Am., 90, S1–S31.
Frankel, A., Mueller, C., Barnhard, T., Perkins, D., Leyendecker, E., Dickman, N., Hanson, S., and
Hopper, M. 1996. “National Seismic Hazard Maps: Documentation June 1996,” U.S. Geological
Survey, Open-File Rept. 96-532.
Frankel, A.D., Mueller, C.S., Barnhard, T.P., Leyendecker, E.V., Wesson, R.L., Harmsen, S.C., Klein,
F.W., Perkins, D.M., Dickman, N.C., Hanson, S.L., and Hopper, M.G. 2000. “USGS National
Seismic Hazard Maps,” Earthquake Spectra, 16, 1–19.
Gupta, A.K. 1993. Response Spectrum Method in Seismic Analysis and Design, CRC Press, Boca Raton, FL.
Hanks, T.C. and Kanamori, H. 1979. “A Moment-Magnitude Scale,” J. Geophys. Res., 84, 2348–2350.
Heaton, T.H., Tajima, F., and Mori, A.W. 1986. “Estimating Ground Motions Using Recorded Accel-
erograms,” Surv. Geophys., 8, 25–83.
Herrmann, R.B. 1985. “An Extension of Random Vibration Theory Estimates of Strong Ground
Motion to Large Distances,” Bull. Seismol. Soc. Am., 75, 1447–1453.
Hough, S.E., Anderson, J.G., Brune, J., Vernon, F., III, Berger, J., Fletcher, J., Haar, L., Hanks, T., and
Baker, L. 1988. “Attenuation Near Anza, California,” Bull. Seismol. Soc. Am., 78, 672–691.
Hwang, H.H.M., Huijie, L., and Huo, J.R. 1997. “Site Coefficients for Design of Buildings in Eastern
North America,” Soil Dyn. Earthquake Eng., 16, 29–40.
ICBO (International Council of Building Officials). 1997. Uniform Building Code, International Coun-
cil of Building Officials, Whittier, CA.
ICC (International Code Council). 2000. International Building Code, International Code Council,
Falls Church, VA.
Johnston, A.C. 1994. “Seismotectonic Interpretations and Conclusions from the Stable Continental
Region Seismicity Database,” in The Earthquakes of Stable Continental Regions, Vol. 1, Assessment
of Large Earthquake Potential, Electric Power Research Institute, Palo Alto, CA, pp. 1–103.
Johnston, A.C. 1996. “Seismic Moment Assessment of Earthquakes in Stable Continental Regions. I.
Instrumental Seismicity,” Geophys. J. Int., 124, 381–414.
Joyner, W.B. 1984. “A Scaling Law for the Spectra of Large Earthquakes,” Bull. Seismol. Soc. Am., 74,
1167–1188.
Joyner, W.B. 2000. “Strong Motion from Surface Waves in Deep Sedimentary Basins,” Bull. Seismol.
Soc. Am., 90, S95–S112.
Joyner, W.B. and Boore, D.M. 1981. “Peak Horizontal Acceleration and Velocity from Strong-Motion
Records Including Records from the 1979 Imperial Valley, California, Earthquake,” Bull. Seis-
mol. Soc. Am., 71, 2011–2038.
Joyner, W.B. and Boore, D.M. 1988. “Measurement, Characterization, and Prediction of Strong
Ground Motion,” in Proc. Conference on Earthquake Engineering and Soil Dynamics II—Recent
Advances in Ground-Motion Evaluation, J.L. von Thun, Ed., June 27–30, 1988, Park City, UT,
Geotech. Spec. Pub. No. 20, pp.43–102, American Society of Civil Engineers, New York.
Joyner, W.B. and Boore, D.M. 1993. “Methods for Regression Analysis of Strong-Motion Data,” Bull.
Seismol. Soc. Am., 83, 469–487.
Joyner, W.B. and Boore, D.M. 1994. “Errata: Methods for Regression Analysis of Strong-Motion
Data,” Bull. Seismol. Soc. Am., 84, 955–956.
Joyner, W.B. and Fumal, T.E. 1984. “Use of Measured Shear-Wave Velocity for Predicting Geologic
Site Effects on Strong Ground Motion,” in Proc. 8th World Conference on Earthquake Engineering,
Vol. 2, July 21–28, San Francisco, pp. 777–783, Prentice-Hall, Englewood Cliffs, NJ.
Joyner, W.B., Warrick, R.E., and Fumal, T.E. 1981. “The Effect of Quaternary Alluvium on Strong
Ground Motion in the Coyote Lake, California, Earthquake of 1979,” Bull. Seismol. Soc. Am.,
71, 1333–1349.
Kanamori, H. 1978. “Quantification of Earthquakes,” Nature, 271, 411–414.
Lay, T. and Wallace, T.C. 1995. Modern Global Seismology, Academic Press, San Diego.
Lee, Y. and Anderson, J.G. 2000. “Potential for Improving Ground-Motion Relations in Southern Cali-
fornia by Incorporating Various Site Parameters,” Bull. Seismol. Soc. Am., 90, S170–S186.
Leyendecker, E.V., Hunt, J.R., Frankel, A.D., and Rukstales, K.S. 2000. “Development of Maximum
Considered Earthquake Ground Motion Maps,” Earthquake Spectra, 16, 21–40.
Liu, L. and Pezeshk, S. 1999. “An Improvement on the Estimation of Pseudoresponse Spectral
Velocity Using RVT Method,” Bull. Seismol. Soc. Am., 89, 1384–1389.
Loh, C.H. 1999. “Interpretation of Structural Damage in 921 Chi-Chi Earthquake,” in Proc. Interna-
tional Workshop on Chi-Chi, Taiwan Earthquake of September 21, 1999, Dec. 14–17, Tai-Chung,
Taiwan, ROC.
Marone, C. and Scholz, C.H. 1988. “The Depth of Seismic Faulting and the Upper Transition from
Stable to Unstable Slip Regimes,” Geophys. Res. Lett., 15, 621–624.
McGarr, A. 1984. “Scaling of Ground Motion Parameters, State of Stress, and Focal Depth,” J.
Geophys. Res., 89, 6969–6979.
Molas, G.L. and Yamazaki, F. 1995. “Attenuation of Earthquake Ground Motion in Japan Including
Deep Focus Events,” Bull. Seismol. Soc. Am., 85, 1343–1358.
Molas, G.L. and Yamazaki, F. 1996. “Attenuation of Response Spectra in Japan Using New JMA
Records,” Bull. Earthquake Resistant Struct. Res. Center, 29, 115–128.
Moores, E.M. and Twiss, R.J. 1995. Tectonics, W.H. Freeman and Company, New York.
Newmark, N.M. and Hall, W.J. 1982. Earthquake Spectra and Design, Earthquake Engineering Research
Institute, Berkeley, CA.
Ou, G.B. and Herrmann, R.B. 1990. “A Statistical Model for Ground Motion Produced by Earth-
quakes at Local and Regional Distances, Bull. Seismol. Soc. Am., 80, 1397–1417.
Park, S. and Elrick, S. 1988. “Predictions of Shear-Wave Velocities in Southern California Using
Surface Geology,” Bull. Seismol. Soc. Am., 88, 677–685.
Richter, C.F. 1935. “An Instrumental Earthquake Magnitude Scale,” Bull. Seismol. Soc. Am., 25, 1–32.
Rodriguez-Marek, A., Bray, J.D., and Abrahamson, N.A. 2001. “An Empirical Geotechnical Seismic
Site Response Procedure,” Earthquake Spectra, 17, 65–87.
Sabetta, F. and Pugliese, A. 1996. “Estimation of Response Spectra and Simulation of Nonstationary
Earthquake Ground Motions,” Bull. Seismol. Soc. Am., 86, 337–352.
Sadigh, K., Chang, C.Y., Abrahamson, N.A., Chiou, S.J., and Power, M.S. 1993. “Specification of Long-
Period Ground Motions: Updated Attenuation Relationships for Rock Site Conditions and
Adjustment Factors for Near-Fault Effects,” in Proc. ATC-17-1 Seminar on Seismic Isolation,
Passive Energy Dissipation, and Active Control, Vol. I, March 11–12, San Francisco, pp. 11–23,
Applied Technology Council, Redwood City, CA.
Sadigh, K., Chang, C.Y., Egan, J.A., Makdisi, F., and Youngs, R.R. 1997. “Attenuation Relationships
for Shallow Crustal Earthquakes Based on California Strong Motion Data,” Seismol. Res. Lett.,
68, 180–189.
Savy, J.B., Bernreuter, D.L., and Chen, J.C. 1987. “A Methodology to Correct for Effect of the Local
Site Characteristics in Seismic Hazard Analyses,” in Ground Motion and Engineering Seismology,
A.S. Cakmak, Ed., Developments in Geotechnical Engineering, Vol. 44, pp. 243–255, Elsevier,
Amsterdam.
Schnabel, P.B. and Seed, H.B. 1973. “Accelerations in Rock for Earthquakes in the Western United
States,” Bull. Seismol. Soc. Am., 63, 501–516.
Schneider, J.F., Silva, W.J., and Stark, C.L. 1993. “Ground Motion Model for the 1989 M 6.9 Loma
Prieta Earthquake Including Effects of Source, Path and Site,” Earthquake Spectra, 9, 251–287.
Si, H. and Midorikawa, S. 2000. “New Attenuation Relations for Peak Ground Acceleration and
Velocity Considering Effects of Fault Type and Site Condition,” in Proc. 12th World Conference
on Earthquake Engineering, Jan. 30–Feb. 4, Auckland, Proc. CD-ROM, Paper 532, New Zealand
Society of Earthquake Engineering, Silverstream, New Zealand.
Silva, W. and Darragh, R.B. 1995. Engineering Characterization of Strong Ground Motion Recorded at
Rock Sites, Rept. EPRI TR-102262, Electric Power Research Institute, Palo Alto, CA.
Silva, W.J. and Lee, K. 1987. “WES RASCAL Code for Synthesizing Earthquake Ground Motions,”
in State-of-the-Art for Assessing Earthquake Hazards in the United States, Paper S-73-1, Rept. 24,
Department of the Army, U.S. Army Corps of Engineers, Vicksburg, MS.
Singh, J.P. 1985. “Earthquake Ground Motions: Implications for Designing Structures and Recon-
ciling Structural Damage,” Earthquake Spectra, 1, 239–270.
Singh, S. and Herrmann, R.B. 1983. “Regionalization of Crustal Coda Q in the Continental United
States,” J. Geophys. Res., 88, 527–538.
Somerville, P. 2000. “New Developments in Seismic Hazard Estimation,” in Proc. 6th International
Conference on Seismic Zonation, Nov. 12–15, Palm Springs, CA, Proc. CD-ROM, Earthquake
Engineering Research Institute, Oakland, CA.
Somerville, P. and Abrahamson, N. 1995. “Ground Motion Prediction for Thrust Earthquakes,” in
Proc. SMIP95 Seminar on Seismological and Engineering Implications of Recent Strong-Motion Data,
May 16, San Francisco, pp. 11–23, California Strong Motion Instrumentation Program, Sac-
ramento, CA.
Somerville, P. and Abrahamson, N. 2000. “Prediction of Ground Motions for Thrust Earthquakes,”
Data utilization report CSMIP/00-01, California Strong Motion Instrumentation Program,
Sacramento, CA.
Somerville, P.G., Smith, N.F., Graves, R.W., and Abrahamson, N.A. 1997. “Modification of Empirical
Strong Ground Motion Attenuation Relations to Include the Amplitude and Duration Effects
of Rupture Directivity,” Seismol. Res. Lett., 68, 199–222.
Somerville, P., Collins, N., Abrahamson, N., Graves, R., and Saikia, C. 2001. “Ground Motion Atten-
uation Relations for the Central and Eastern United States, U.S. Geological Survey, Award
99HQGR0098, Final Report.
Spudich, P., Joyner, W.B., Lindh, A.G., Boore, D.M., Margaris, B.M., and Fletcher, J.B. 1999. “SEA99:
A Revised Ground Motion Prediction Relation for Use in Extensional Tectonic Regimes,” Bull.
Seismol. Soc. Am., 89, 1156–1170.
Stewart, J.P. 2000. “Variations between Foundation-Level and Free-Field Earthquake Ground
Motions,” Earthquake Spectra, 16, 511–532.
Toro, G.R., Abrahamson, N.A., and Schneider, J.F. 1997. “Model of Strong Ground Motions from
Earthquakes in Central and Eastern North America: Best Estimates and Uncertainties,” Seismol.
Res. Lett., 68, 41–57.
Trifunac, M.D. and Lee, V.W. 1979. Dependence of Pseudo-Relative Velocity Spectra of Strong Motion
Acceleration on the Depth of Sedimentary Deposits, Rept. CE79-02, Dept. of Civil Engineering,
University of Southern California, Los Angeles.
Tsai, Y.B. and Huang, M.W. 1999. Strong Ground Motion Characteristics of the Chi-Chi, Taiwan Earth-
quake of September 21, 1999, Institute of Geophysics, National Central University, Chung-Li,
Taiwan, ROC.
Tsai, Y.B. and Huang, M.W. 2000. “Strong Ground Motion Characteristics of the Chi-Chi, Taiwan
Earthquake of September 21, 1999,” Earthquake Eng. Eng. Seismol., 2, 1–21.
Wells, D.L. and Coppersmith, K.J. 1994. “New Empirical Relationships among Magnitude, Rupture
Length, Rupture Width, Rupture Area, and Surface Displacement,” Bull. Seismol. Soc. Am., 84,
974–1002.
Westaway, R. and Smith, R.B. 1989. “Strong Ground Motion in Normal-Faulting Earthquakes,”
Geophys. J., 96, 529–559.
Wills, C.J. and Silva, W. 1998. “Shear-Wave Velocity Characteristics of Geologic Units in California,”
Earthquake Spectra, 14, 533–556.
Wills, C.J., Petersen, M., Bryant, W.A., Reichle, M., Saucedo, G.J., Tan, S., Taylor, G, and Treiman, J.
2000. “A Site-Conditions Map for California Based on Geology and Shear-Wave Velocity,”
Bull. Seismol. Soc. Am., 90, S187–S208.
Youngs, R.R., Abrahamson, N.A., Makdisi, F., and Sadigh, K. 1995. “Magnitude Dependent Disper-
sion in Peak Ground Acceleration,” Bull. Seismol. Soc. Am., 85, 1161–1176.
Youngs, R.R., Chiou, S.J., Silva, W.J., and Humphrey, J.R. 1997. “Strong Ground Motion Attenuation
Relationships for Subduction Zone Earthquakes,” Seismol. Res. Lett., 68, 58–73.
Zoback, M.L. 1992. “First- and Second-Order Patterns of Stress in the Lithosphere: The World Stress
Map Project,” J. Geophys. Res., 97, 11,703–11,728.
Further Reading
There have been many reviews and discussions of attenuation relations and their use in engineering.
A recent comprehensive review is provided in “Earthquake Ground Motion Estimation Using
Strong-Motion Records: A Review of Equations for the Estimation of Peak Ground Acceleration
and Response Spectral Ordinates,” by J. Douglas in the journal Earth-Science Reviews.
On the related but broader topic of engineering seismology, there are several good summaries including
“Strong Motion Seismology,” by John Anderson in the International Handbook of Earthquake Engineering
and Seismology; “Seismology, Engineering,” by Kenneth Campbell in the Encyclopedia of Physical Science
and Technology, 3rd ed.; and Engineering Seismology, by Kiyoshi Kanai.
Recent advances in the prediction of strong ground motion are published in several earthquake
engineering and seismology journals, including the Bulletin of the Seismological Society of America and
Seismological Research Letters, both published by the Seismological Society of America; Earthquake
Spectra, published by the Earthquake Engineering Research Institute; Earthquake Engineering and
Structural Dynamics, published by John Wiley & Sons; and Soil Dynamics and Earthquake Engineering,
published by Elsevier Science Publishers.
Other useful reviews include:
Abrahamson, N.A. and Shedlock, K.M. 1997. “Overview,” Seismol. Res. Lett., 68, 9–23.
Ambraseys, N.N. and Bommer, J.J. 1995. “Attenuation Relations for Use in Europe: An Overview,”
in Proc. 5th SECED Conference on European Seismic Design Practice, E.A. Elnashai, Ed., October
26–27, Chester, United Kingdom, pp. 67–74, A.A. Balkema, Rotterdam.
Boore, D.M. 1987. “The Prediction of Strong Ground Motion,” in Strong Ground Motion Seismology,
M.O. Erdik and M.N. Toksoz, Eds., D. Reidel, Dordrecht, the Netherlands, pp. 109–141.
Boore, D.M. and Joyner, W.B. 1982. “The Empirical Prediction of Ground Motion,” Bull. Seismol.
Soc. Am., 72, S43–S60.
Campbell, K.W. 1985. “Strong Motion Attenuation Relations: A Ten-Year Perspective,” Earthquake
Spectra, 1, 759–804.
Iai, S. and Brady, A.G. 1993. Proceedings of the International Workshop on Strong Motion Data, Dec.
13–17, Menlo Park, CA, Vols. 1 and 2, The Port and Harbour Research Institute, Kanagawa,
Japan.
Idriss, I.M. 1979. “Characteristics of Earthquake Ground Motions,” in Proc. ASCE Specialty Conference
on Earthquake Engineering and Soil Dynamics, Vol. III, June 19–21, Pasadena, CA., pp. 1154–1261,
American Society of Civil Engineers, New York.
Idriss, I.M. 1993. Procedures for Selecting Earthquake Ground Motions at Rock Sites, Rept. NIST GCR 93-
625, National Institute of Standards and Technology, Gaithersburg, MD.
Idriss, I.M. 1995. “An Overview of Earthquake Ground Motions Pertinent to Seismic Zonation,” in
Proc. 5th International Conference on Seismic Zonation, Vol. III, Oct. 17–19, Nice, France, pp.
2111–2126, Ouest Editions, Presses Academiques, Nantes Cedex, France.
Joyner, W.B. and Boore, D.M. 1988. “Measurement, Characterization, and Prediction of Strong
Ground Motion,” in Proc. Conference on Earthquake Engineering and Soil Dynamics II—Recent
Advances in Ground-Motion Evaluation, J.L. von Thun, Ed., June 27–30, Park City, UT, Geotech.
Spec. Pub. No. 20, pp.43–102, American Society of Civil Engineers, New York.
Joyner, W.B. and Boore, D.M. 1990. “Empirical Methods of Ground Motion Estimation,” in Proc.
POLA Seismic Workshop on Seismic Engineering, R.C. Wittkop and G.R. Martin, Eds., March
21–23, San Pedro, CA, pp. 273–308, Port of Los Angeles.
Joyner, W.B. and Boore, D.M. 1996. “Recent Developments in Strong Motion Attenuation Relation-
ships,” in Proc. 28th Joint Meeting of the U.S.–Japan Cooperative Program in Natural Resource Panel
on Wind and Seismic Effects, N.J. Raufaste, Ed., May 14–17, Gaithersburg, MD, pp. 101–116,
National Institute of Standards and Technology, Gaithersburg, MD.
Seed, H.B. and Idriss, I.M. 1982. Ground Motions and Soil Liquefaction during Earthquakes, Earthquake
Engineering Research Institute, Berkeley, CA.
Somerville, P. 1999. “Recent Advances in Strong Ground Motion Prediction,” in Proc. 8th Canadian
Conference on Earthquake Engineering, June 13–16, Vancouver, Canada, pp. 7–28, Canadian Asso-
ciation for Earthquake Engineering, Vancouver.
Somerville, P. 2000. “New Developments in Seismic Hazard Estimation,” in Proc. 6th International Con-
ference on Seismic Zonation, Nov. 12–15, Palm Springs, CA, Proc. CD-ROM, Earthquake Engineer-
ing Research Institute, Oakland, CA.
Appendix–Notation
Depth Parameters
drup Average depth to top of the rupture plane (km)
dseis Average depth to top of the seismogenic part of the rupture plane (km)
hhypo Hypocentral depth (also focal depth) (km)
hrup Depth to the point on the fault rupture plane that corresponds to rrup (km)
Hbot Depth to the bottom of the seismogenic part of the fault (km)
Htop Depth to the top of the fault (km)
Hseis Depth to the top of the seismogenic part of the fault (km)
Miscellaneous Parameters
f Wave frequency (1/T, Hz)
fn Natural frequency of a single-degree-of-freedom system (1/Tn , Hz)
g Fraction of gravity (980.6550 cm/sec2)
T Period (1/f, sec)
Tn Natural period of a single-degree-of-freedom system (1/fn , sec)
W Down-dip width of the fault rupture plane (km)
zT Indicator variable for subduction interface and intraslab events
δ Angle of the fault plane with respect to the Earth’s surface (dip angle)
π Archimedes number (3.141593)