Attenuation Relationships for Shallow Crustal
Earthquakes Based on California Strong
Motion Data
K. Sadigh
C.-Y. Chang
J.A. Egan
F. Makdisi
R.R. Youngs
Geomatrix Consultants
ABSTRACT
Attenuation relationships are presented for peak acceleration
and response spectral accelerations from shallow crustal
earthquakes. The relationships are based on strong motion
data primarily from California earthquakes. Relationships
are presented for strike-slip and reverse-faulting earthquakes,
rock and deep firm soil deposits, earthquakes of moment
magnitude M 4 to 8+, and distances up to 100 km.
INTRODUCTION
In this paper we summarize attenuation relationships we
have developed from the analysis of strong motion data
recorded primarily in California. These relationships have
evolved through several iterations as new data have been
gathered. The starting point is the set of attenuation rela-
tionships for peak acceleration and 5 percent damped spec-
tral accelerations for rock and soil sites presented by Sadigh
etaL (1986). Subsequently, Sadigh etal. (1989, 1993) devel-
oped an updated version of the rock site attenuation rela-
tionships that included analyses of strong motion data from
the 1989 M 7 Lorna Prieta, 1992 M 7.3 Landers, and M 6.4
Big Bear earthquakes. In this paper we present an updated
version of the soil site attenuation relationships that also
includes analysis of strong motion data from the 1994 M 6. 7
Northridge earthquake, together with rock site attenuation
models.
STRONG MOTION DATA
The ground motion data on which the attenuation relation-
ships are based consist primarily of accelerograms from Cal-
ifornia earthquakes of moment magnitude M 3.8 and greater
recorded at sites within 200 km of the rupture surface. Data
180 Seismological Research Letters Volume 68, Number 1 January/February 1997
from the Gazli, USSR (1976) and Tabas, Iran (1978) earth-
quakes were also included to provide more large magnitude
data. Table 1lists the earthquakes that have been considered
in the development of these relationships. Figure 1 shows the
distributions of these events in terms of magnitude and
source to site distance. We characterize the earthquake size
by moment magnitude, M, as defined by Hanks and Kan-
amori (1979). Distance is defined as the minimum distance
to the rupture surface, rrup· However, for many of the smaller
magnitude events, rupture surfaces have not been defined
and we use hypocentral distance. Because the dimensions of
rupture for small events are usually much smaller than the
distances to the recording stations, we do not believe that the
use of hypocentral distance introduces significant bias into
the attenuation models. We distinguish between strike-slip
and reverse-faulting earthquakes by the rake angle of rup-
ture, with rake angles greater than 45 o considered reverse-
faulting events, and rake angles less than 45 o considered
strike-slip events. Examination of the peak motion data from
the small number of normal-faulting earthquakes in the data
set indicated that they were not significantly different from
peak motions from strike-slip earthquakes. Therefore, the
normal and strike-slip earthquakes were combined into a
single category. Peak horizontal acceleration (PGA) and
response spectral acceleration (SA) are represented by the
geometric mean of the two horizontal components.
Attenuation relationships are presented for two general
site categories, rock and deep soil. Rock sites are those with
bedrock within about a meter from the surface. Recent stud-
ies of the shear wave velocities at many strong motion sta-
tions that have been classified in the past as rock indicate that
the surface velocity often falls below the classical rock defini-
tion of >750 m/sec, and there is usually a strong velocity gra-
dient because of near-surface weathering and fracturing.
Thus, the site conditions representative of the rock attenua-
 TABLE 1
List of Earthquakes Used to Develop Attenuation Relationships
No. of Records2
Distance Range
Earthquake Date M Fault Type 1 (km) R DS
Kern County,CA 1952/07/21 7.4 RV 120.5-224.0 0 3
Port Hueneme, CA 1957/03/18 4.7 RV 14.1-14.1 0 1
Daly City, CA 1957/03/22 5.3 RV 9.5-9.5 1 0
Parkfield, CA 1966/06/27 6.1 ss 0.1-230.0 1 6
Borrego Mtn., CA 1968/04/09 6.6 ss 113.Q-261.0 5 3
Santa Rosa, CA (A) 1969/10/02 5.6 ss 80.Q-113.0 1 2
Santa Rosa, CA (B) 1969/10/02 5.7 ss 78.9-112.0 1 2
Lytle Creek, CA 1970/09/12 5.3 RV 19.7-76.0 5 2
San Fernando, CA 1971/02/09 6.6 RV 2.8-305.0 11 14
Lake Isabella, CA 1971/03/08 4.1 ss 8.9-8.9 1 0
Bear Valley, CA 1972/02/24 4.7 ss 2.5-2.5 1 0
Point Mugu, CA 1973/02/21 5.6 RV 25.0-25.0 0 1
Hollister, CA 1974/11/28 5.2 ss 39.0-39.0 1 0
Oroville, CA 1975/08/01 5.9 ss 9.5-35.8 2 2
Oroville, CA (R) 1975/08/02 5.1 ss 12.7-14.6 0 2
Oroville, CA (S) 1975/08/02 5.2 ss 12.4-15.0 0 2
Oroville, CA (A) 1975/08/03 4.6 ss 8.4-14.9 1 6
Oroville, CA (B) 1975/08/03 4.1 ss 7.7-15.0 1 6
Oroville, CA (F) 1975/08/06 4.7 ss 10.9-16.1 1 7
Oroville, CA (K) 1975/08/08 4.9 ss 6.5-13.3 2 5
Oroville, CA (N) 1975/08/11 4.3 ss 2.5-11.6 2 4
Oroville, CA (P} 1975/08/16 4.0 ss 9.5-12.3 3 5
Oroville, CA (T} 1975/09/26 4.0 ss 10.8-19.9 3 5
Oroville, CA (U} 1975/09/27 4.6 ss 8.6-22.4 3 5
Gazli, USSR 1976/05/17 6.8 RV 3.0-3.0 1 0
Calipatria Swarm,CA 1976/11/04 4.9 ss 13.6-26.4 1 2
Santa Barbara, CA 1978/08/13 6.0 RV 18.5-21.0 2 1
Tabas, Iran 1978/09/16 7.4 RV 17.Q-172.0 3 0
Coyote Lake, CA 1979/08/06 5.7 ss 3.1-63.3 4 7
Imperial Valley, CA 1979/10/15 6.5 ss 0.5-158.0 4 33
lmp.Val., CA (A02) 1979/10/15 3.8 ss 11.6-32.2 0 6
lmp.Val., CA (A03) 1979/10/15 5.2 ss 12.2-27.1 0 15
lmp.Val., CA (A05} 1979/10/15 4.0 ss 9.7-19.5 0 2
lmp.Val., CA (A07} 1979/10/15 4.2 ss 14.4-24.9 0 3
lmp.Val., CA (A10} 1979/10/15 4.2 ss 10.6-21.6 0 3
lmp.Val., CA (A13} 1979/10/15 4.6 ss 9.6-10.2 0 2
lmp.Val., CA (A15} 1979/10/15 4.3 ss 11.9-24.6 0 3
lmp.Val., CA (A16) 1979/10/15 4.0 ss 5.2-5.2 0 1
lmp.Val., CA (A21) 1979/10/15 4.5 ss 10.4-21.2 0 2
lmp.Val., CA (A22) 1979/10/15 4.5 RV 23.6-27.6 1 1
lmp.Val., CA (A25) 1979/10/15 5.1 RV 14.8-33.8 0 9
lmp.Val., CA (A26) 1979/10/15 4.0 ss 12.1-13.9 0 2
lmp.Val., CA (A27) 1979/10/15 4.1 ss 13.2-13.2 0 1

Seismological Research Letters Volume 68, Number 1 January/February 1997 181

 TABLE 1 (Continued)
List of Earthquakes Used to Develop Attenuation Relationships
No. of Records2
Distance Range
Earthquake Date M Fault Type 1 (km) R DS
lmp.Val., CA (A29) 1979/10/15 5.1 ss 14.1-33.5 0 9
lmp.Val., CA (A30) 1979/10/15 4.6 ss 16.6-16.6 0 1
lmp.Val., CA (A31) 1979/10/15 5.5 ss 16.1-39.7 1 10
lmp.Val., CA (A32) 1979/10/16 4.2 ss 11.3-11.3 0 1
lmp.Val., CA (A34) 1979/10/16 4.0 RV 14.3-14.3 0 1
I
lmp.Val., CA (A35) 1979/10/16 4.8 ss 17.6-29.6 0 3
lmp.Val., CA (A36) 1979/10/16 4.0 ss 15.0-20.1 0 2
lmp.Val., CA (A38) 1979/10/16 4.9 ss 23.Q-25.9 0 3
lmp.Val., CA (A43) 1979/10/17 4.1 ss 19.4-19.4 0 1
lmp.Val., CA (A44) 1979/10/17 4.5 ss 7.7-19.6 0 4
lmp.Val., CA (A51) 1979/12/21 4.6 ss 6.3-21.3 0 4
Livermore, CA 1980/01/24 5.8 ss 20.3-67.0 4 3
Livermore, CA 1980/01/26 5.4 ss 8.0-62.1 5 3
Horse Canyon, CA 1980/02/25 5.3 ss 5.8-60.0 3 9
Mammoth Lakes, CA (A) 1980/05/25 6.2 ss 15.5-42.4 3 1
Mammoth Lakes (B) 1980/05/25 5.7 ss 24.3-24.3 2 0
Mammoth Lakes (C) 1980/05/25 6.0 ss 19.7-19.7 3 0
Mammoth Lakes (C1) 1980/05/25 5.7 ss 14.4-14.4 3 0
Mammoth Lakes (0006) 1980/05/27 4.9 ss 17.1-48.6 4 6
Mammoth Lakes (0009) 1980/05/27 3.9 ss 5.5-5.5 0 1
Mammoth Lakes (0010) 1980/05/27 3.9 ss 9.9-9.9 0 1
Mammoth Lakes (0011) 1980/05/28 3.8 ss 9.6-15.7 0 2
Mammoth Lakes (0012) 1980/05/28 3.8 ss 6.4-11.3 0 3
Mammoth Lakes (0014) 1980/05/28 4.8 ss 7.9-9.4 0 2
Mammoth Lakes (0018) 1980/05/28 4.2 ss 18.9-26.1 0 2
Mammoth Lakes (0021) 1980/05/29 4.0 ss 14.7-19.9 0 2
Mammoth Lakes (0026) 1980/05/29 4.0 ss 19.1-19.1 0 1
Mammoth Lakes (0038) 1980/05/30 3.8 ss 9.5-11.2 0 2
Mammoth Lakes (0040) 1980/05/31 4.6 ss 16.6-23.7 0 3
Mammoth Lakes (0042) 1980/05/31 4.1 ss 5.1-11.6 0 3
Mammoth Lakes (0050) 1980/05/31 4.9 ss 7.7-11.8 0 3
Mammoth Lakes (0051) 1980/05/31 3.9 ss 8.Q-12.1 0 2
Mammoth Lakes (0055) 1980/05/31 3.9 ss 11.5-14.1 0 2
Mammoth Lakes (0065) 1980/06/02 4.1 ss 9.6-14.0 0 3
Mammoth Lakes (0067) 1980/06/02 3.9 ss 8.7-10.7 0 2
Mammoth Lakes (0089) 1980/06/05 4.3 ss 11.5-15.9 0 2
Mammoth Lakes (0112) 1980/06/07 3.9 ss 5.2-8.6 0 3
Mammoth Lakes (0113) 1980/06/08 3.9 ss 12.5-18.6 0 3
Mammoth Lakes (0115) 1980/06/08 4.4 ss 25.7-25.7 0 1
Mexicali Valley, Mex. 1980/06/09 6.4 ss 5.7-33.3 1 5
Westmorland, CA 1981/04/26 5.6 ss 10.1-69.8 1 22
Coalinga, CA 1983/05/02 6.5 RV 24.6-74.5 12 29
Coalinga, CA (A02) 1983/05/06 3.9 ss 9.2-9.2 1 0

182 Seismological Research Letters Volume 68, Number 1 January/February 1997

 TABLE 1 (Continued)
List of Earthquakes Used to Develop Attenuation Relationships
No. of Records2
Distance Range
Earthquake Date M Fault Type 1 (km) R OS
Coalinga, CA (A03) 1983/05/08 5.1 RV 12.1-17.8 3 3
Coalinga, CA (A04) 1983/05/08 4.5 RV 12.5-12.5 1 0
Coalinga, CA (ADS) 1983/05/12 4.4 RV 13.9-13.9 0 1
Coalinga, CA (A07) 1983/05/24 4.6 RV 8.9-16.1 2 1
Coalinga, CA (A08) 1983/06/10 5.3 RV 9.7-10.5 2 2
Coalinga, CA (A09) 1983/06/11 4.0 RV 19.9-19.9 1 0
Coalinga, CA (A 10) 1983/07/09 5.2 RV 10.4-17.0 4 2
Coalinga, CA (A 11) 1983/07/18 4.2 RV 13.7-13.7 0 1
Coalinga, CA (A12) 1983/07/21 5.9 RV 9.5-15.3 4 2
Coalinga, CA (A 13) 1983/07/21 4.9 RV 10.2-13.7 4 2
Coalinga, CA (A14) 1983/07/25 5.2 RV 10.0-14.7 4 2
Coalinga, CA (A 16) 1983/09/09 5.3 ss 8.2-18.4 2 1
Coalinga, CA (A17) 1983/09/11 4.5 RV 14.5-16.8 1 2
Morgan Hill, CA 1984/04/24 6.2 ss 0.1-157.9 13 24
Bishop, CA 1984/11/23 5.8 ss 13.3-43.1 1 4
Hollister, CA 1986/01/26 5.4 ss 16.9-38.1 3 7
N. Palm Spr., CA 1986/07/08 6.1 ss 7.3-154.0 15 20
Chalf.Val., CA (FS) 1986/07/20 5.9 ss 11.0-27.0 3 4
Chalf.Val., CA 1986/07/21 6.3 ss 8.5-40.6 4 4
Chalf.Val., CA (A1) 1986/07/21 5.6 ss 13.0-20.0 1 3
Chalf.Val., CA (A2) 1986/07/31 5.8 ss 12.0-21.0 0 3
Cerro Prieto, Mex 1986/11/30 5.4 ss 5.6-13.4 1 2
Whitt.Nar., CA 1987/10/01 6.0 RV 17.2-112.0 18 29
Whitt.Nar., CA (A) 1987/10/04 5.3 ss 19.9-56.9 5 6
Sup.Hills, CA (A) 1987/11/23 6.3 ss 12.5-126.9 1 26
Sup.Hills, CA (B) 1987/11/24 6.7 ss 0.1-135.0 4 41
Sup.Hills, CA (B1) 1987/11/27 4.2 ss 19.5-19.5 1 0
Sup.Hills, CA (B2) 1987112/01 4.6 ss 22.6-22.6 1 0
Whitt.Nar., CA (A2) 1988/02/11 5.0 ss 17.Q-17.0 0 1
Lorna Prieta, CA 1989/10/17 7.0 ss 5.1-122.7 26 32
Upland, CA 1990/02/28 5.5 ss 11.2-13.0 0 2
Sierra Madre, CA 1991/06/28 5.5 RV 12.6-76.9 8 20
Landers, CA 1992/06/28 7.3 ss 1.1-175.6 20 65
Big Bear, CA 1992/06/28 6.4 ss 10.0-50.4 0 7
Northridge, CA 1994/01/17 6.7 RV 4.1-64.8 16 35
1. RV =Rake > 45', SS = Rake < 45'
2. R =Rock, DS =deep soil

Seismological Research Letters Volume 68, Number 1 January/February 1997 183

 Strike Slip Reverse
•
... at• CII*i_W_ 0 . tl 0
• 0 Occ.m'l> 1

-~c:aJO·--- 0
• 0 ct
0 •

• •
• ·~
0
o+

+ Rock + Rock
0
•
0 Deep Soil 0 Deep Soil

3
2 5 10 20 50 100 200300 1 2 5 10 20 50 100 200300

Distance (km) Distance (km)

.A Figure 1. Distribution of strong motion data used in development of attenuation relationships.

tion models given here should be considered soft rock. The
deep soil data are from sites with greater than 20 m of soil
over bedrock. We have not included data from very soft soil
sites, such as those on San Francisco bay mud. Thus, the rela-
tionships are considered appropriate for deep, firm soil
deposits. The data used in the analyses are from free-field
recordings, that is from instruments housed in instrument
shelters located near the ground surface and from the ground
floor of small, light structures.
ATTENUATION MODEL DEVELOPMENT
Historically there are typically more data for peak accelera-
tion than for response spectral acceleration, and the set of
digitized and processed accelerograms tends to be the larger
amplitude recordings from any individual earthquake.
Therefore, the process that we have used to develop the
attenuation relationships consists of two stages. First, atten-
uation relationships are developed for PGA by regression
analyses using the general form
where ZT is an indicator variable taking the value 1 for
reverse events and 0 for strike slip events. We have found the
need to develop different coefficients for events larger and
smaller than M "" 6Vz to account for near-field saturation
effects.
In the second stage of the analysis, relationships for
spectral amplification (SNPGA) are fit to the response spec-
tral ordinate data normalized by the PGA of the recordings.
The form that we have found to work well is:
In addition, Sadigh eta!. (1993) added a term to account for
near-field high-frequency motion (coefficient C7 in Table 2).
The final attenuation models for SA are obtained by
combining (1) and (2). The resulting parameters were then
smoothed to produce attenuation relationships that predict
smooth response spectra over the full range of magnitudes
(M 4 to 8+) and distances (rrup 0 to 100 km).
Once the median relationships were defined, the data-
base was used to compute standard errors for PGA and SA at
individual periods. The standard errors were found to be
dependent on magnitude (Youngs et aL, 1995) and were rep-
resented by linear relationships between magnitude and
standard error. The relationships for individual periods were
smoothed to produce smooth estimates of 84th percentile
response spectra.
The attenuation models that we have developed are
listed in Tables 2 through 4 for rock and deep soil site condi-
tions, respectively. Figures 2 and 3 show comparisons of the
median attenuation relationships and recorded PGA data for
strike slip and reverse earthquakes, respectively. Figure 4
shows the predicted median spectral shapes for rock and
deep soil site motions.

184 Seismological Research Letters Volume 68, Number 1 January/February 1997

 TABLE 2
Attenuation Relationships of Horizontal Response Spectral Accelerations (5% Damping) for Rock Sites
+ C 4 In(rrup + exp( C 5 + C6 M)) + C7 In(rrup + 2)
25
ln(y) = C1 + C2 M + C 3 (8.5M) '

Period(s) c1 c2 c3 c4 Cs Cs c7
ForM:::; 6.5
PGA -0.624 1.0 0.000 -2.100 1.29649 0.250 0.0
0.07 0.110 1.0 0.006 -2.128 1.29649 0.250 -o.082
0.10 0.275 1.0 0.006 -2.148 1.29649 0.250 -0.041
0.20 0.153 1.0 -o.004 -2.080 1.29649 0.250 0.0
0.30 -0.057 1.0 -o.017 -2.028 1.29649 0.250 0.0
0.40 -0.298 1.0 -Q.028 -1.990 1.29649 0.250 0.0
0.50 -0.588 1.0 -o.040 -1.945 1.29649 0.250 0.0
0.75 -1.208 1.0 -o.o5o -1.865 1.29649 0.250 0.0
1.00 -1.705 1.0 -Q.055 -1.800 1.29649 0.250 0.0
1.50 -2.407 1.0 -0.065 -1.725 1.29649 0.250 0.0
2.00 -2.945 1.0 -o.070 -1.670 1.29649 0.250 0.0
3.00 -3.700 1.0 -o.o8o -1.610 1.29649 0.250 0.0
4.00 -4.230 1.0 -Q.100 -1.570 1.29649 0.250 0.0
ForM> 6.5
PGA -1.274 1.1 0.000 -2.100 -0.48451 0.524 0.0
0.07 -0.540 1.1 0.006 -2.128 -0.48451 0.524 -0.082
0.10 -0.375 1.1 0.006 -2.148 -0.48451 0.524 -0.041
0.20 -D.497 1.1 -o.004 -2.080 -0.48451 0.524 0.0
0.30 -0.707 1.1 -0.017 -2.028 -0.48451 0.524 0.0
0.40 -0.948 1.1 -0.028 -1.990 -0.48451 0.524 0.0
0.50 -1.238 1.1 -o.040 -1.945 -0.48451 0.524 0.0
0.75 -1.858 1.1 -o.o5o -1.865 -0.48451 0.524 0.0
1.00 -2.355 1.1 -o.055 -1.800 -0.48451 0.524 0.0
1.50 -3.057 1.1 -o.065 -1.725 -0.48451 0.524 0.0
2.00 -3.595 1.1 -o.070 -1.670 -0.48451 0.524 0.0
3.00 -4.350 1.1 -o.o8o -1.610 -0.48451 0.524 0.0
4.00 -4.880 1.1 -Q.100 -1.570 -0.48451 0.524 0.0
Note: Relationships for reverse/thrust faulting are obtained by multiplying the above strike-slip amplitudes by 1.2.

Seismological Research Letters Volume 68, Number 1 January/February 1997 185

DISCUSSION TABLE 3
Dispersion Relationships for Horizontal Rock Motion
The attenuation relationships listed in Tables 2 through 4 are
considered applicable for estimating free field ground
Period Sigma [ln(y)J
motions from shallow crustal earthquakes in the magnitude PGA 1.39- 0.14M; 0.38 for ~ 7.21
range of M 4 to 8+. Shallow crustal earthquakes are those 0.07 1.40 - 0.14M; 0.39 for ~ 7.21
that occur on faults within the upper 20 to 25 km of conti-
0.10 1.41-0.14M;0.40for~7.21
nental crust. The relationships were developed for reverse
and strike-slip faulting earthquakes. We did not find a signif- 0.20 1.43- 0.14M; 0.42 for ~ 7.21
icant difference between peak motions from strike-slip earth- 0.30 1.45- 0.14M; 0.44 for ~ 7.21
quakes and the limited number of data from normal faulting 0.40 1.48- 0.14M; 0.47 for~ 7.21
earthquakes and combined these two types of earthquakes
0.50 1.50- 0.14M; 0.49 for~ 7.21
into a single category. The data used to develop the relation-
ships are primarily from California and occur in both com- 0.75 1.52- 0.14M; 0.51 for~ 7.21
pressional and extensional stress regimes. Campbell (1987) 1.00 1.53- 0.14M, 0.52 for~ 7.21
found no significant difference between the peak motions for >1.00 1.53- 0.14M; 0.52 for ~ 7.21
events occurring in compressional and extensional stress
regimes. motions are high, the soil motions become lower than rock
The attenuation relationships show an expected trend motions presumably because of nonlinear site response
for soil versus rock motions, i.e., soil amplitudes are larger effects. The rock motion spectral shapes change with dis-
than rock where the rock motions are low because of sire tance. However, no significant effect of distance was
amplification in the lower velocity soil layers. Where the rock observed in the soil spectral shapes within 50 km of the

TABLE 4
Attenuation Relationship Coefficients for Deep Soil Sites

ln(y) = C1 +C2 M-C ln(rrup +C ec M)+C

3 4
5
6 +C7 (8.5- Mt 5

where y is spectral accleration in g

cl = -2.17 for strike-slip, -1.92 for reverse and thrust earthquakes
C2 = 1.0
c3= 1.7o
C4 = 2.1863, C5 = 0.32 forM$. 6.5
C4 = 0.3825, C5 = 0.5882 forM> 6.5
rrup = closest distance to rupture surface
Period (sec) C6 Strike-Slip C6 Reverse c7 Standard Error1
PGA 0.0 0.0 0.0 1.52- 0.16M
0.075 0.4572 0.4572 0.005 1.54-0.16M
0.1 0.6395 0.6395 0.005 1.54- 0.16M
0.2 0.9187 0.9187 -o.004 1.565- 0.16M
0.3 0.9547 0.9547 -0.014 1.58- 0.16M
0.4 0.9251 0.9005 -o.024 1.595- 0.16M
0.5 0.8494 0.8285 -o.033 1.61- 0.16M
0.75 0.7010 0.6802 -o.051 1.635- 0.16M
1.0 0.5665 0.5075 -o.065 1.66- 0.16M
1.5 0.3235 0.2215 -0.090 1.69- 0.16M
2.0 0.1001 -o.0526 -0.108 1.70- 0.16M
3.0 -Q.2801 -Q.4905 -D.139 1.71-0.16M
4.0 -o.6274 -o.8907 -Q.160 1.71-0.16M
1. Standard error for magnitudes greater than M 7 set equal to the value for M 7

186 Seismological Research Letters Volume 68, Number 1 January/February 1997

 2

i l. ir r
'

l •
t
1
.5 0

.2
-... .1
tr>
,_...
.05
23 oo
~\ ] [ ~0() ~ J L <J
:~
0
~

r r
~--~ 0
1-

\
.02 -t 1- 0 0 8
0
<0

l \, ~ r
Q
0 ~ Strike Slip

\J
SiriKe SHp o roo 0 Strike Slip
C/)
.0.
.005 M 3.8-4.Z '\
Skiko SHp
M 4.3-4.5 M 4.8-5.2 M 5.3-5.5
(T)
c;;· Rock • R~;~ck t Rock
t t Rock ' t
3
0
0
<C
.002 o Deep soil \ o Deep soil o Deep soil f o Deep soil

c:;· .001
~
:0 2
(T)
fJ)
(T)
tl) 1
......
0
=r .5
I
(T)
::+
(T)
......
fJ)
.2
-... .1
< tr>
,_...
£
c: .05
3(T)
a>
00.
23
Q, .02
z
c:
.01 ~

·~~ ~
3

~f
Strike Slip Strike Slip Strike Slip Strike Slip

1[
C"
...,
(T)
.005 M 5.8-6.2 M 6.3-6.5 M 6.8-7.2 M 7.3-7.5

c....
tl)
::l
.002 t__:_::
t Rock
soil
t Rock
o Deep soli
t Rock
o Deep soli
t Rock
o Deep soil
c: .001
tl)

'?. 2 5 10 20 50 100 300 1 2 5 10 20 50 100 300 1 2 5 10 20 50 100 300 1 2 5 10 20 50 100 300

,
(T)
...,
C"
c:
Distance (km) Distance (km) Distance (km) Distance (km)
tl)

-<
~

tO
tO
-..! • Figure 2. Comparison of PGA data for strike-slip (rake < 45°) earthquakes with median attenuation relationships listed in Tables 2 through 4. Solid line is the rock motion rela-
~
tionship from Table 2 and dashed line is the deep soil relationship from Table 4.
00
-..!
CXl
CXl 2
en
CD
u;·
3 .5
0
0
<C
•
c;· .2
~
::0 ...........
D")
.1 •
CD
en
CD
.._., t>
....,
i:l> .05
•
(.")
::::;
r-
0
Q, .02
CD
:::+
....,
CD
.01
en

~
Reverse Reverse Reverse
< .005 M 3.8-4.2 M 4.3-4.5 M 5.3-5.5
2. t Rock t Rock t Rock
c:
3 .002
CD
(j)
CXl .001
2
c: 2
3
0' •
....,
CD
......
I F •
.5
c....
i:l>

~
:::l
c: .2
i:l>

~
"T1
........... .1
CD D")
..._...,
....,
0'
c: .05
....,
i:l>
2§
'<
...... Q, .02
co
co
'-!
.01

~
Reverse Reverse Reverse
.005 M 5.8-6.2 M 6.3-6.5 M 6.8-7.2 M 7.3-7.5
• Rock • Rock • Rock • Rock
.002

.001
t o Deep soli o Deep soil o Deep soil o Deep soil

2 5 10 20 50 100 300 1 2 5 10 20 50 100 300 1 2 5 10 20 50 100 300 1 2 5 10 20 50 100 300

Distance (km) Distance (km) Distance (km) Distance (km)

.._ Figure 3. Comparison of PGA data for reverse (rake> 45·) earthquakes with median attenuation relationships listed in Tables 2 through 4. Solid line is the rock motion relationship
from Table 2 and dashed line is the deep soil relationship from Table 4.
 --- Rock, strike slip and reverse Deep soil, strike slip - 'It Deep soil, reverse
5

~
~ .5

U3
.2
1t

.1
M 6.5, 10 km
.05
5

~
2 # ~

:\
~ •
~ .5

U3
.2

.1
M 5.5, 50 km M 6.5, 50 km M 7.5, 50 km
.05
.02 .05 .1 .2 .5 2 5 .02 .05 .1 .2 .5 2 5 .02 .05 .1 .2 .5 2 5

Period (sec) Period (sec) Period (sec)

.A Figure 4. Comparison of spectral shapes (SA!PGA) obtained using the median attenuation relationships listed in Tables 2 through 4.

rupture. Sadigh et al. (1993) indicate that within 10 km of
the rupture surface there are systematic differences between
the fault-normal and fault-parallel components of long-
period ground motion. Specifically, they recommend that
the fault-normal component be increased by 20 percent over
the geometric average values for spectral periods of 2.0 sec-
onds and greater and the fault-parallel component would be
expected to be 20 percent lower than the geometric average
value. II
REFERENCES
Campbell, K.W. (1987). Predicting strong ground motion in Utah, in
Assessment of &gional Earthquake Hazards and Risk Along the
Wasatch Front, Utah, U.S. Geological Survey Open-File Repon
87-585, II, L-1-120.
Hanks, T.C., and H. Kanamori (1979). A moment magnitude scale,]
Geophys. Res., 84, 2,348-2,350.
Sadigh, K., C.-Y. Chang, N.A. Abrahamson, S.J. Chiou and M.S.
Power ( 199 3). 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 Semi-
nar on Seismic Isolation, Passive Energy Dissipation, and Active
Control, March 11-12, San Francisco, California, 59-70.
Sadigh, K., C.-Y. Chang, F. Makdisi and J.A Egan, (1989). Attenua-
tion relationships for horizontal peak ground acceleration and
response spectral acceleration fur rock sites (abs.), Seism. &s. Lett.,
60, 19.
Sadigh, K., J.A. Egan, and R.R. Youngs (1986). Specification of
ground motion for seismic design oflong period structures (abs.),
Earthquake Notes, 57, n. 1, 13. Relationships printed in W.B.
Joyner and D.M. Boore (1988), Measurement, characterization,
and prediction of strong ground motion, in Earthquake Engineer-
ing and Soil Dynamics !/-&cent Advances in Ground Motion
Evaluation, ASCE Geotechnical Special Publication 20,43-102.
Youngs, R.R., N.A. Abrahamson, F. Makdisi, and K. Sadigh (1995).
Magnitude dependent dispersion in peak ground acceleration,
Bull. Seism. Soc. Am., 85, 1,161-1,176.
Geomatrix Consultants
100 Pine St., 1Oth Floor
San Francisco, CA 94111

Seismological Research Letters Volume 68, Number 1 January/February 1997 189