Bulletin of the Seismological Society of America, Vol. 70, No. 1, pp.

305-321, February 1980

PEAK ACCELERATION, VELOCITY, AND DI SPLACEMENT FROM
STRONG-MOTION RECORDS
BY DAVID M. BOORE, WILLIAM B. JOYNER, ADOLPH A. OLIVER, III,
AND ROBERT A. PAGE
ABSTRACT
Strong-motion data from earthquakes of western North America are examined
to provide the basis for estimating peak acceleration, velocity, and displacement
as a function of distance for three magnitude classes, 5. 0 to 5.7, 6. 0 to 6.4, and
7.1 to 7.6. Analysis of a subset of the data from the San Fernando earthquake
shows that small but statistically significant differences exist between peak
values of horizontal acceleration, velocity, and displacement recorded on soil at
the base of small structures and values recorded at the base of large structures.
The peak acceleration tends to be less and the peak velocity and displacement
to be greater at the base of large structures than at the base of small structures.
In the distance range used in the regression analysis (15 to 100 km), the values
of peak horizontal acceleration recorded at soil sites in the San Fernando
earthquake are not significantly different from the values recorded at rock sites,
but values of peak horizontal velocity and displacement are significantly greater
at sol
/ I INTRODUCTION
/
Peak horizontal acceleratior, is commonly used to scale response spectra or
ground-motion/time histories for use in earthquake resistant design, particularly for
nuclear poSeUr plant facilities (Newmarket al , 1973). Methods have been proposed
(Newmark and Hall, 1969) for constructing design spectra using three peak param-
et ers-hori zont al acceleration, velocity, and displacement--the advantage of using
all three parameters being that together they convey some information concerning
the shape of the spectrum as well as the amplitude level. In this paper, an analysis
of a large number of earthquake data is presented as a basis for estimating the peak
acceleration, velocity, and displacement for a hypothetical earthquake of a pre-
scribed magnitude at a prescribed distance from the causative fault. This work is a
continuation of that reported by Page et al. (1972) and Page et al. (1975); it is a
condensation of U.S. Geological Survey Circular 795 by Boore et al. (1978),
hereafter referred to as USGS Circular 795, which should be consulted for further
details.
It is not our purpose to advocate the use of peak parameters in scaling design
motions. We look forward to the development of new methods for prescribing design
motions, methods more firmly based in the physics that govern faulting and wave
propagation. Pending the development of such methods, we recognize widespread
current practice and at t empt to present the available strong-motion data in a
compact and useful form for estimating peak parameters.
DATA CHARACTERISTICS AND METHODS OF PRESENTATION
Sources of data. The data set includes 204 recordings from 19 earthquakes of
western North America, taken primarily from volumes I and II of the series "Strong-
motion Earthquake Accelerograms" published by the Earthquake Engineering
Research Laboratory of the California Institute of Technology. The data set used is
tabulated in USGS Circular 795; a listing of the events used is given in Table 1.
3o5
306 D. M. BOORE, W, B. J OYNER, A. A. OLI VER, AND R. A. PAGE
Di st ances. In all cases, t he di st ance used is t he short est di st ance bet ween t he
surface of faul t slippage and t he recordi ng point. Thi s would clearly be t he preferred
measure of di st ance if equal areas of t he faul t emi t t ed equal amount s of seismic
energy and if t he surface were known. The second condi t i on is somet i mes not met ;
t he first is probabl y never met . Ot her measures of di st ance have been used in
analysis of st rong-mot i on data, part i cul arl y epi cent ral distance, hypocent r al dis-
tance, and di st ance from t he cent er of energy rel ease (Duke et al., 1972). The use of
epi cent ral di st ance or hypocent r al di st ance has t he advant age t hat t hese measures
are mor e commonl y known and special studies are not requi red to det er mi ne t hem.
In some cases, however, t hese measures are clearly i nappropri at e, as for a long faul t
r upt ur e wi t h t he epi cent ral at one end and recordi ng st at i ons at t he ot her. The use
TABLE 1
EARTHQUAKES PROVIDING DATA USED IN THIS PAPER
Date Origin Time
(y/m/d) (h:min UTC) Location Magnitude
72/02/24 15:56 Bear Valley, California 5.0
74/11/28 23:01 Bear Valley, California 5.2
75/06/07 08:46 Ferndale, California 5.2
57/03/22 19:44 Daly City, California 5.3
70/09/12 14:30 Lytle Creek, California 5.4
66/06/28 04:26 Parkfield, California 5.5
67/06/21 18:04 Fairbanks, Alaska 5.6
69/10/02 04:56 Santa Rosa, California 5.6
69/10/02 06:19 Santa Rosa, California 5.7
75/08/01 20:20 Oroville, California 5.7
73/02/21 14:45 Point Mugu, California 6.0
72/12/23 06:29 Managua, Nicaragua 6.2
40/05/19 04:36 Imperial Valley, California 6.4
68/04/09 02:28 Borrego Mtn., California 6.4
71/02/09 14:00 San Fernando, California 6.4
49/04/13 19:55 Puget Sound, Washington 7.1
59/08/18 06:37 Hebgen Lake, Montana 7.1
52/07/21 I1:52 Kern County, California 7.2
72/07/30 21:45 Sitka, Alaska 7.6
of di st ance t o t he cent er of energy release is a way of avoiding t he assumpt i on of
uni form radi at i on over t he r upt ur e surface, but for long r upt ur es this measure, too,
may be i nappropri at e. In our opi ni on t he best choice for general purposes is t he
closest di st ance t o t he r upt ur e surface, but t he uncert ai nt i es resulting from nonuni -
f or m radi at i on over t he surface shoul d be kept in mind. An illustration of t hose
uncert ai nt i es is provi ded by t he Pacoi ma Dam recordi ng of t he San Fer nando
ear t hquake of 1971. By i nt erpret at i ons of a number of seismological dat a (e.g.,
Hanks, 1974; Boor e and Zoback, 1974), t he sources for t he peak vel oci t y and t he
peak accel erat i on on t he record are t hought to come f r om different regions of t he
faul t separat ed by per haps 20 km; nei t her region cont ai ns t he closest poi nt t o t he
i nst rument .
In order t o avoi d obscuri ng t he at t enuat i on relation, we generally exclude dat a
where t he uncer t ai nt y in di st ance is large. Following Page et al. (1972), t he distances
are classified as A, B, or C, accordi ng t o t he uncer t ai nt y (less t han 2 km, 2 t o 5 kin,
and 5 to 25 km, respectively). C-quality dat a are used onl y for t he magni t ude 7.1
Puget Sound ear t hquake and t he magni t ude 7.1 Hebgen Lake eart hquake. In t he
pl ot s t o follow, t he uncer t ai nt y class A, B, or C is i ndi cat ed by t he size of t he symbol,
t he l argest for class A and t he smallest for class C.
ESTIMATING PEAK ACCELERATION, VELOCITY, AND DISPLACEMENT 307
Cl as s i f i cat i on o f dat a. The dat a have been divided into classes according to
magnitude, site geology, and size of associ at ed structure. The dat a were divided into
t hree magni t ude classes (5.0 to 5.7, 6.0 to 6.4, and 7.1 to 7.6) on t he basis of t he
Ri cht er local magni t ude (Richter, 1958), if available; ot herwi se surface-wave mag-
ni t ude was used. (Sources of dat a for assigning magni t udes are included in US GS
Ci r cul ar 795.) The Imperial Valley ear t hquake was assigned a magni t ude of 6.4 in
accordance with a det ermi nat i on by Tri funac and Br une (1970) and in cont rast to
t he value 7.1 t hat is commonl y given.
Kanamori and Jenni ngs (1978) have recent l y devel oped a met hod of det ermi ni ng
Ri cht er local magni t ude from st rong-mot i on records. Thei r magni t ude assi gnment s
are in general agreement wi t h ours. The largest difference is for t he Puget Sound
ear t hquake of 1949, for which their val ue is 6.5 and our val ue is 7.1.
Recordi ng sites are assigned to one of two categories, "rock" or "soil", by applying
our best j udgment to t he available site descriptions. St at i ons are assigned to t he
rock cat egory if t hey are underlain by mat eri al descri bed by such t erms as granite,
diorite, gneiss, chert, graywacke, limestone, sandstone, siltstone, or shale. St at i ons
are assigned to t he soil cat egory if t hey are underlain by a sufficient t hi ckness of
mat eri al descri bed by such t erms as alluvium, sand, gravel, clay, silt, mud, fill, or
glacial outwash. If, from t he site description, soil mat eri al overlying rock is j udged
to be less t han 4 to 5 m thick, t he site is assigned to t he rock category. (Sources for
site descriptions are given in US GS Ci r cul ar 795.) Because considerable uncert ai nt y
and ambi gui t y at t ends t he geological classification of recording sites, conclusions
t hat rely on t he validity of t he classification of a single st at i on are omi t t ed. Only
t rends reveal ed by comparing whole classes of dat a are considered.
Many of t he dat a come from t he basement or ground floor of buildings or from
t he abut ment s of dams. In t he analysis of st rong-mot i on data, it is commonl y
assumed t hat (he influence of t he st ruct ure on t he mot i on of t he base can be ignored
and t hat t he dat a as recorded represent free-field ground motion. A limited t est of
this assumpt i on has been at t empt ed by classifying recording sites in accordance
wi t h t he size of t he associ at ed st ruct ure, one class for sites at t he base of one- t o
t wo-st ory buildings and anot her for sites at t he base of taller buildings or on dam
abut ment s. Compari son of t he two classes using dat a from t he San Fer nando
ear t hquake is descri bed in a subsequent section.
One woul d expect t he vel oci t y and di spl acement dat a from small st ruct ures to be
more represent at i ve of free-field motion. The transfer functions relating mot i on at
t he base of st ruct ures t o free-field mot i on t end t oward uni t y for frequenci es t hat are
small relative to t he fixed-base nat ural frequencies of t he structure. (For examples
of t heoret i cal and empirical t ransfer functions, see Duke et al., 1970; Crouse and
Jennings, 1975.) Small st ruct ures generally have nat ural frequencies in t he range 2
t o 10 Hz, whi ch iz significantly above t he range of frequenci es domi nant in t he
vel oci t y and di spl acement t i me histories. The case of acceleration is more compli-
cated. For large buildings, t he domi nant acceleration frequencies will be higher t han
t he fundament al structtiral resonant frequencies, and here t he t ransfer functions
t end to fall bel ow unity. The nat ural frequenci es of t he small st ruct ures, however,
are in t he same range as t he frequencies domi nant in t he acceleration t i me histories,
and t he effect of t he st ruct ure may be to raise or lower t he peak acceleration,
dependi ng on t he spect r um of t he eart hquake and t he details of t he transfer function.
We woul d expect t he acceleration val ues for t he large st ruct ures to be syst emat i cal l y
bi ased downward, but t he values for t he smaller st ruct ures may be ei t her increased
or decreased. In fact, compari son of San Fernando dat a shows smaller accel erat i ons
on t he average for t he large st ruct ures. Therefore, focus is pl aced on t he dat a from
308 D. M. BOORB, W. B. JOYNER, A. A. OLI VER, AND R. A. PAGE
t he small st ruct ures as a basis for estimating free-field motion.
Geographic distribution. In an at t empt t o keep t he dat a sampl e reasonabl y
homogeneous, only records obt ai ned in t he west ern par t of Nor t h America were
included. In order to avoid bi as from t he ext remel y dense cluster of i nst rument s in
downt own Los Angeles, a special selection procedure was used in t he area bet ween
l at i t ude 34.00 and 34.11N and longitude 118.24 and 118.45W. Wi t hi n each of t he
two geological site categories, only one recording per ear t hquake was allowed for
each st ruct ure category, making a maxi mum of four possible recordings from t he
desi gnat ed area for one eart hquake. Sel ect i on was made by choosing t he st at i on
wi t h t he smal l est identification number of all eligible stations.
Presentation of data. Peak horizontal acceleration, velocity, and di spl acement are
pl ot t ed against distance on log-log grids for each magni t ude class. The peak val ues
for horizontal mot i on are t aken from t he component wi t h t he larger peak.
Statistics. The nat ure of t he st rong-mot i on dat a set is not such as t o bear t he
wei ght of el aborat e or subt l e statistical inferences. For t hat reason, plots showing
t he individual dat a poi nt s are emphasized. We do, however, indulge in statistical
analysis to t he ext ent of fitting l east -squares straight lines relating t he l ogari t hm of
t he peak par amet er s t o t he l ogari t hm of distance and det ermi ni ng t he confidence
limits for t he predi ct i on of a single value of t he dependent variable (Dixon and
Massey, 1957); t he 70 and 95 per cent predi ct i on intervals are shown by t he pairs of
curved lines. The equat i ons used in t he statistical analysis and t he coefficients for
t he regression lines shown in t he pl ot s are cont ai ned in USGS Circular 795. An
at t empt has been made t o avoid bi as in t he regression analysis by not including
poi nt s t hat are ei t her t oo close or t oo far from t he fault. For close points, t he dat a
are t oo sparse t o indicate t he proper functional form for t he regression; for far
points, t he dat a set is i ncompl et e because not all i nst rument s were triggered by t he
motion. The straight lines clearly fit t he dat a as well as woul d any ot her simple
relation. Curvat ure t hat mi ght be caused by anelastic at t enuat i on is compl et el y
obscured by t he scat t er in t he data.
The scat t er is approxi mat el y const ant i ndependent of distance. Thi s const ancy
suggests t hat it was correct t o fit a straight-line relation t o t he logarithms of
vari abl es r at her t han fit a power-law relation t o t he vari abl es t hemsel ves.
ANALYSI S OF THE WHOLE DATA S ET
Dat a from horizontal component s for all t he ear t hquakes are present ed in this
section, using t he dat a from small st ruct ures because, for t he reasons given previ-
ously, t hose dat a are considered t o be a bet t er guide to free-field motion. (Dat a from
larger buildings and vertical component dat a are cont ai ned in USGS Circular 795.)
Coefficients from t he regression anal yses are cont ai ned in Tabl e 2.
Horizontal acceleration. Peak horizontal accel erat i on dat a for t he t hree magni-
t ude classes (Figure 1) show t hat accel erat i on clearly increases wi t h magni t ude in
t hose distance ranges for whi ch t here is overl ap of dat a bet ween t he magni t ude
classes. The scat t er for t he magni t ude 5.0 to 5.7 dat a is significantly great er t han
t hat for ei t her of t he ot her two classes. Thi s difference may arise part l y because a
number of different ear t hquakes cont ri but e subst ant i al l y to t he dat a set for t he 5.0
to 5.7 class, whereas t he 6.0 to 6.4 class is domi nat ed by dat a from t he 1971 San
Fer nando ear t hquake and t he 7.1 t o 7.6 class is domi nat ed by dat a from t he 1952
Ker n Count y eart hquake.
The rat e of at t enuat i on of accel erat i on wi t h distance for t he magni t ude 5.0 t o 5.7
ESTIMATING PEAK ACCELERATION, VELOCITY, AND DISPLACEMENT 309
class appears to be greater t han indicated by the slope of -0. 9 for the mean
regression line in Figure 1. This is suggested by the systematic tendency for the data
points at distances beyond 30 km to lie below an extension of the mean regression
line. As previously explained, we have chosen to exclude from the regression analysis
data beyond the distance at which all instruments can be presumed to have
triggered. The distance range for which a reasonably complete data set is available
is not adequate for a good determination of slope; the standard error of the slope for
the magnitude 5.0 to 5.7 class is 0.5. Judging from t he data at greater distances, t he
slope of -1.2 + 0.3 for the mean line for the magnitude 6.0 - 6.4 class appears to be
a better estimate of the rate of attenuation to distances of at least 100 km for t hat
data set. The slope of -2. 0 + 0.4 for the magnitude 7.1 to 7.6 class may overesti mate
the rate of attenuation, but the data are scanty.
Horizontal velocity. The peak horizontal velocity data for the three magnitude
classes are presented in Figure 2. There are fewer points of velocity t han acceleration
TABLE 2
STATISTICAL PARAMETERS OF LoG-LOG ATTENUATION RELATIONS*
Distance Range b :i:
Magnitude Range (km) a Standard Error Data Pointa
Accel er at i on (Fi gure 1)
5.0 t o 5.7 5 to 30 0.2 - 0. 9 + 0.5 19
6.0 to 6.4 15 to 55 1.0 - 1. 2 + 0.3 16
7.1 to 7.6 40 to 150 2.6 - 2. 0 + 0.4 9
Vel oci t y (Fi gure 2)
5.3 to 5.7 5 to 30 2.4 - 1. 2 +_ 0.6 11
6.4 15 to 55 1.9 - 0. 6 _+ 0.4 14
Di spl acement {Figure 3)
5.3 to 5.7 5 to 30 1.8 - 1. 2 + 0.6 11
6.4 15 to 55 1.5 - 0. 6 + 0.5 14
* For m of regressi on equat i on: l ogl 0(ampl i t ude) = a - b log~o(distance).
because integrations were not available for all the accelerograms. There are so few
points for the magnitude 7.1 to 7.2 class t hat regression lines are not included. As
with acceleration, the peak velocity at a given distance tends to increase with
magnitude.
The slope of -0.6 _+ 0.4 for the mean regression lines for the magnitude 6.4 data
appears to underestimate the rate of attenuation if one considers the San Fernando
data {described in the next section), which give better determinations because the
distance range extends to 100 km. We were confident t hat all the instruments out to
100 km were triggered in the San Fernando earthquake, but this confidence does
not apply to the whole magnitude class.
Horizontal displacement. The peak horizontal displacements for the three mag-
nitude classes are given in Figure 3. The scatter of the data is larger than for
acceleration or velocity in each magnitude class, and the standard errors of the
slopes of the mean regression lines exceed 0.5. The displacements are derived from
double integration of high-pass filtered accelerograms and therefore represent high-
pass filtered versions of the true ground displacement. The longer periods, which
are contaminated by processing noise, are removed.
Hanks (1975), in studying the errors in displacement records derived by double
310
C~
z
~E
O
I
I.--
0.1
_.1
L.LJ
u
u
7

N 0. 01
r v
O
0. 001
D. M. BOORE, W. B. JOYNER , A. A. OLIVER~ AND R. A. PAGE
MAGNI TUDE 5. 0 - 5. 7 SMALL STRUCTURE
I [ I I I ~ I I I I [ F I ~- I I
\
0:
EXPLANATI ON X 0
X
X Rock si t e
O Soi l si t e ~
I ~ I I l l l ~l I I I I ] l l l l I ~
10 100
DISTANCE, IN KILOMETERS
MAGNI TU DE 6. 0- 6. 4 SMALL STRUCTU RES
[ I I I I I I I I I 1 I l l l l l I I I
1. 0'
O'1
z__ o e
g
~ 0.1
g
N g go
0. 001 I I I I I l l i I I l l l l l l I 1
10 100
DISTANCE, IN KILOMETERS
O0
Z
z
O
0. 1
o
u
0
S
o.oi
0
"I -
MAGNI TUDE 7. 1- 7. 6 SMALL STRUCTURES 70-PERCENT PREDICTION INTERVALS
].o - ~.o 5.0 - 5.7 _
~ 6. 0- 6.4
: \ ,X 7.6
i \1 \
o g ,,
) . 1
<~
0 ~) 001
0.001 I I I I I I I l l I L I Jl l l l l I I ~ 0. 1p01 I I I I I I l l l I [ I I I I I l l I I
10 100 10 100
DISTANCE, IN KILOMETERS DISTANCE, IN KILOMETERS
Fro. 1. Peak horizontal acceleration recorded at base of small structures versus distance to slipped
fault for three magnitude ranges. Center line, mean regression line. Outer pair of lines represents 95 per
cent prediction interval; inner pair, 70 per cent prediction interval. Length of lines represents distance
interval considered in regression analysis. Uncertainty in distance is inversely related to symbol size (see
text). Lower right figure shows a comparison of the 70 per cent prediction intervals for the three
magnitude ranges.
integration of filtered accelerograms, found that the errors are typically less than 1
cm in the period range 5 to 8 sec, 1 to 2 cm at periods near 10 sec, and 2 to 4 cm in
the period range 10 to 15 sec. These findings raise the possibility that some of the
l ow-ampl i tude data poi nts in Figure 3 may be i nfl uenced by noi se and may represent
ESTIMATING PEAK ACCELERATION, VELOCITY, AND DISPLACEMENT 311
MAGNITUDE 5.3 - 5.7 ,SMALL STRUCTURES
r ~
Z
O
U100. 0
rv
I.-- 10. 0
Z
M.I
u
z
O
._.1
~ 1.0
g
z
O
N
O
r i i p i l i t ~ [ i i i i : T
)<
EXPLANATI ON
x Rock si t e
O Soi l si t e
I I
Z
O
U 100. 0
r v
r v
10. 0
Z
ii,iII
D
Z
i
>i-
I
(,D
0
,i-ii ~.o
i:-

N
O
112 0.1
MAGNITUDE 6.4 SMALL STRUCTURES
I I I I I I I I I I I I l l l l l I [ J
X
I I I I t l i l d I I ] i i i 1[ I I I I I I I I I I l [ [ I [ l l l l l [ I J
10 100 10 100
DISTANCE, IN KILOMETERS DISTANCE, IN KILOMETERS
MAGNITUDE 7. 1- 7. 2 SMALL STRUCTURES
I I I I ] l l l I l I I I Fl l l 1 I I
Z
0
,i(J 100. 0-
i ' v
~-- O
t -- 10. 0
Z ~ @
I.l.I
u
z
>i.-
i i
I
u
0
1.o
z
O
N
O
- r
0.1
Z
0
,,u,
c, O
~0
Z
Z
0
o
0
N
0
"1-
100. 0
10. 0
1.0
70-PERCENT PREDICTION INTERVALS
I I F I I l l l I I [ [ I l l l l E [ I
5.3 -5. 7
~, xT. 1- 7. 2
q
I I I I I I I I I I I I I I P I I 0.1 I t I I r l l ~l I I I I I I l l
10 100 10 100
DISTANCE, IN KILOMETERS DISTANCE, IN KILOMETERS
1
Fro. 2. Peak hori zont al velocity. See Figure 1 for explanation of symbols and curves. Dashed curves
in the lower right figure for class 7.1 to 7.2 emphasize uncertainty in slope.
upper bounds to the actual ground displacement. The character of some of the low
amplitude records resembles noise rather than signal. Nevertheless, we have pro-
ceeded in the analysis with the understanding that the results may be compromised
to some extent by the effect of noise on the weaker motions.
The peak displacement at a given distance increases with magnitude in a manner
similar to that observed in the acceleration and velocity data.
312
r- v
Z ]o.o
u
7
Z
u
<
u) 1.0 ~
Z
o
0
- r
D. M. BOORE, W. B. JOYNER, A. A. OLIVER, AND R. A. PAGE
MAGNITUDE 5.3-5.7 SMALL STRUCTURES
\ \
\
EYSo il sile
0.1
O
X
~_l l l l l i i i I I I L2/ ___L I
10 loo
DISTANCE, IN KILOMETERS
Z lO.O
u
z
~.~ 1.0
Z
0
N
0
0.1
MAGNITUDE 6.4 SMALL STRUCTURES
? ~ I I I l i i l j I I I ql i l l I I I
1 I I I I I I l l I I I I l t l l l I I
10 100
DISTANCE, IN KILOMETERS
MAGNITUDE 7.1- 7. 2 SMALL STRUCTURES
~_ I I I I 11111 I i i qqqT~- - T~m~
l,o
~'---~ 10.0~- 0 0
0
N
7
N
o
0.1
70-PERCENT PREDICTION INTERVALS
I I ] l l l l l I ~ I I l i l l L I I I
5. 3- 5. 7
7. 1- 7. 2
\6<:,,
\L "-.< \ ,
\ I 7 \
~ \ kt I
I " I
\ \
\
~7 10.0 --
U
z
Z
u..i
u
<
~ 1.0
N
0
N
N

I
0.1
10 100 10 100
DISTANCE, IN KILOMETERS DISTANCE, IN KILOMETERS
FI G. 3. Peak horizontal displacement. See Figure 1 for explanation of symbols and curves.
ANALYSIS OF DATA FROM THE SAN FERNANDO EARTHQUAKE
The San Fernando earthquake supplied more than one-quarter of the total data
points in our sample. The large number of data points from a single event provides
the best basis for examining the effects of structures and local site conditions. The
San Fernando earthquake also gives more accurate values than the whole magnitude
6.0 to 6.4 data set for the slopes of the regression lines for peak parameters against
ESTIMATING PEAK ACCELERATION, VELOCITY, AND DISPLACEMENT 313
d i s t a n c e . T h i s a c c u r a c y i s p o s s i b l e b e c a u s e , a s me n t i o n e d p r e v i o u s l y , t h e s t a t i s t i c a l
a n a l y s i s c a n b e c a r r i e d o u t o v e r a g r e a t e r r a n g e of d i s t a n c e f or t h e S a n F e r n a n d o
e a r t h q u a k e . As d i s c u s s e d e a r l i e r , t o a v o i d b i a s n o t a l l t h e r e c o r d s f r o m d o wn t o wn
Lo s An g e l e s a r e i n c l u d e d i n t h e d a t a s et . T h e c o e f f i c i e n t s f r o m t h e r e g r e s s i o n
a n a l y s e s a r e g i v e n i n T a b l e s 3 a n d 4; f or mo r e d e t a i l s e e USGS Circular 795.
I n c o mp a r i n g p e a k p a r a me t e r s f or d i f f e r e n t s i z e s o f s t r u c t u r e s a n d s i t e c o n d i t i o n s ,
TABLE 3
STATISTICAL PARAMETERS OF LoG-Lo6 ATTENUATION RELATIONS
FOR THE SAN FERNANDO EARTHQUAKE DATA*
b
[dentification a Data Points
Standard Error
Acceleration
Small structures 1.1 -1. 3 + 0.2 12
Large structures 0.9 -1. 3 _+ 0.2 18
Velocity
Small structures 3.1 -1. 3 + 0.2 11
Large structures 2.6 -1. 0 + 0.1 18
Displacement
Small structures 2.1 - 0. 9 + 0.4 11
Large structures 2.1 -0. 8 + 0.2 18
* Soil sites, comparison of recordings at base of small and large
structures (Figure 4).
TABLE 4
STATISTICAL PARAMETERS OF Lo~-Loo REGRESSION RELATIONS FOR
THE SAN FERNANDO EARTHQUAKE DATA*
b
Identification a Data Points
Standard Error
Acceleration
Rock sites 1.4 - 1. 6 + 0.2 l0
Soil sites 1.1 -1. 3 _+ 0.2 12
Velocity
Rock sites 3.1 - 1. 5 _+ 0.4 9
Soil sites 3.1 -1. 3 + 0.2 11
Displacement
Rock sites 2.7 -1. 5 _+ 0.;4 9
Soil sites 2.1 - 0. 9 _+ 0.4 11
* Small structures, comparison of recordings on rock and soil {Figure
5).
a n a n a l y s i s - o f - v a r i a n c e t e c h n i q u e i s u s e d ( Ac t o n , 1959, p. 80- 83) t o t e s t t h e s t a t i s t i c a l
s i g n i f i c a n c e o f t h e r e d u c t i o n i n v a r i a n c e t h a t o c c u r s wh e n d i f f e r e n t r e g r e s s i o n l i ne s
a r e f i t t o t h e t wo d i f f e r e n t d a t a s e t s . I n g e n e r a l , t h e a n a l y s i s - o f - v a r i a n c e t e s t s e n a b l e
us t o s e e h o w t h e d i f f e r e n c e s b e t we e n d a t a s e t s c o mp a r e wi t h t h o s e t h a t mi g h t b e
c a u s e d b y r a n d o m s a mp l i n g e r r o r . T h e r e s h o u l d b e n o s e n s e of c o n f i d e n c e , h o we v e r ,
t h a t t h e s t r o n g - mo t i o n d a t a s e t s r e p r e s e n t r a n d o m s a mp l e s ; i n a n y cas e, t h e
s t a t i s t i c a l t e s t s s a y n o t h i n g a b o u t t h e r e a l p h y s i c a l me a n i n g o f t h e d i f f e r e n c e s
b e t we e n d a t a s e t s .
Effect of structure. F i g u r e 4 c o mp a r e s p e a k hol ~. , zont al a c c e l e r a t i o n , v e l o c i t y , a n d
d i s p l a c e me n t v a l u e s r e c o r d e d o n s oi l a t t h e b a s e o f s ma l l s t r u c t u r e s ( one - a n d t wo-
314 D. M. BOORE, W. B. JOYNER, A. A. OLI VER, AND R. A. PAGE
Z
m ] . 0--
z"
O
t--
o
o
~: 0.1
7
0
S

T
o.o1
u
~- 1QO. O - -
(D
Z
>i-
(.i)
0
10.0
Z

N
(3
1.0
u.I
b-
~u~ 10. 0 - -
z
Z
k)
N
Z
0
U
N
0
"1-
0. 1! I
SAN FERNANDO SOIL SITES SAN FERNANDO SOIL SITES
- - T- ~] ~l 11111 I I I LI I I I j I I I I [ I I I I i I I I I I I
I I I [ l l l l L I
@
I r . l l ~l , ~ i , ~ I ,
O~
@
EXPLANATION
@ Small structure
Large structure
I I
r
- - ~r - I i ~ il I L I I LI I I I
] I [ I I I I ] I 1 I Ell111
\ \ - .
EXPLANATION ~
Small structures
- - ~ Large structures
I I I I I EI I I I I I 11111 I I I I ql l l l l ~ . l I I I l l l
1o 100 10 100
DISTANCE, IN KILOMETERS DISTANCE, IN KILOMETERS
I
I I I
i
i
L I
I
FIG. 4. (Left) Peak mot i ons ver sus di st ance to sl i pped faul t at soil si t es in San Fer nando ear t hquake.
(Right) Compar i son of mean regressi on l i nes and 70 per cent predi ct i on i nt erval s for smal l st r uct ur es
(solid lines) and l arge st r uct ur es ( dashed lines).
ESTI MATI NG PEAK ACCELERATI ON, VELOCITY, AND DI SPLACEMENT 315
story buildings) and large structures (buildings with more t han two stories, or dams).
For acceleration, the mean regression line for data from small structures lies above
t hat for large structures, and the analysis-of-variance tests indicate t hat the differ-
ence is significant at the 90 per cent level. The difference in slope is not significant.
For velocity, the mean regression line for small structures generally lies below t hat
for large structures, although the lines cross. The difference in level is statistically
significant at the 98 per cent level, although unimpressive to the eye. The small-
structures line is steeper, and the difference in slope is significant at the 90 per cent
level. For horizontal displacement, the mean regression line for small structures lies
below t hat for large structures, and the difference is significant at the 99 per cent
level. The difference in slope is not significant.
In summary, for most of the distance range covered by the regression analysis
peak horizontal acceleration is less and peak horizontal velocity and displacement
are greater, on the average, at the base of large structures than at the base of small
structures. The attenuation with distance is greater for the small structures for all
three parameters, but the difference is statistically significant only for peak velocity.
The result t hat acceleration values from the large structures are lower on the
average is expected from considerations of soil-structure interaction and is consistent
with the findings of Crouse (1978). The data shown in Figure 4 encourage us in our
preference for the data from small structures as a basis for estimating free-field
ground motion. In general, however, the differences between the data from large
and small structures are relatively small compared with the range of either data set,
and we do not believe t hat firm conclusions are warranted solely on the basis of
formal statistical tests. The differences may be caused by soil-structure interaction,
but more study would be required to confirm this.
Effect of site geology. Figure 5 compares peak horizontal acceleration, velocity,
and displacement recorded at the base of small structures on rock and soil. For
acceleration, the analysis-of-variance tests indicate t hat the differences in the
regression lines are not significant in either slope or level. For velocity the mean
regression line is higher for soil, and the difference is significant at the 98 per cent
level. The difference in slope is not significant. For displacement the mean regression
line for soil is higher; the difference is significant at the 98 per cent level. The
difference in slope is not significant even at the 75 per cent level.
Peak horizontal acceleration appears to be nearly the same, on the average, on
rock and soil sites, whereas both peak horizontal, velocity and displacement are
larger on soil sites. This relation is not the result of any obvious bias in the data. No
gross effect is evident from bias in the distribution of stations with distance. To test
for bias due to the nonuniform azimuthal distribution of the data (Hanks, 1975), the
azimuth of each station was determined with respect to a point in the center of the
zone of fault rupture (latitude 34.37N, longitude 118.42W). A mean regression line
against distance was determined for all the peak-horizontal-motion data for small
structures in the distance range 15 to 100 km (with distance measured to the closest
point on the rupture surface, as before). Residuals to the regression line are plotted
against azimuth in a polar diagram (Figure 6), the circle representing zero residual.
For acceleration data, no strong systematic difference is apparent between rock and
soil. Although the azimuthal coverage is far from complete, it can be said t hat in
any range of azimuth for which both rock and soil points are present, the soil
residuals for velocity and displacement data are more positive.
It is tentatively concluded t hat amplification of velocity and displacement is a
real effect associated with soil sites; presumably, for the soil sites, some sort of
316 D. M. BOORI% W. B. JOYNER, A. A. OLI VER s AND R. A. PAGE
Z 1.0
z"
0
i i -
<
r ~
u
u
0A
p-
z
0
N
0
"1-
0. 0 I
U
I O0.O
u
z
i.-
U
0
SAN FERNANDO SMALL STRUCTURES SAN FERNANDO SMALL STRUCTURES
i I I I I l l I I I 111111 I I I I I I l l I I I I I I I I I I I
xx
XO00
0 0
@0
>
I 0,0
Z

U
0
"t"
1. 0
F--
Z 10.0
(D
Z
z
U
<
z
0
N

T
- - ~ [ t t L I I T ] l l I T~- -
X
x%
X
X 0 o
X O0 0
X X
x6o x>o
X
I I I I I I l I I l l l l l I I
<>
xx o
0
o<>
x
X
X 0
EXPLANATI ON
X Rock si t e
0 SOil si t e
E7 I t I [ I
~ - [ I l l l l m
I [ I 1411[ I
T i I i 11111
EXPLANATI ON
Rock si t e
~ Soi l si t e
\
X\
X
\
\
0. 1 _. ~- L I I 11111] I I 1111111 I I I I I t l i l l l ! I I I I 1111 l
10 100 10 100
DI STANCE, IN KILOMETERS DISTANCE, IN KILOMETERS
Fro. 5. (Left) Peak mot i ons recorded at base of smal l structures versus di stance to sl i pped fault in
San Fernando earthquake. (Right) Mean regressi on l i nes and 70 per cent predi cti on intervals.
ESTIMATING PEAK ACCELERATION, VELOCITY, AND DISPLACEMENT
SAN FERNANDO SAN FERNANDO
SMALL STRUCTURES ACCELERATION SMALL STRUCTURES VELOCI TY
317
A FACTOR OF TWO


A FACTOROFTWO
SAN FERNANDO
SMALL STRUCTURES DISPLACEMENT

0
A FACTOROFTWO
EXPLANATION
X Rock site
~> Soil site
Fro. 6. Azimuthal dependence of residuals from mean regression lines for acceleration, velocity, and
displacement. X and diamond represent rock and soil sites, respectively.
ampl i fi cat i on mechani sms are oper at i ng on t he l onger peri ods t hat are domi nant on
vel oci t y and di spl acement records. For t he shor t er peri ods domi nant on accel erat i on
records, t hese mechani s ms are count er bal anced by anel ast i c at t enuat i on. Ther e will
be no specul at i on her e on t he nat ur e of t he ampl i fi cat i on mechani sms. Si mi l ar
concl usi ons on t he effect of site condi t i ons on st rong mot i on in t he San Fer nando
ear t hquake were r epor t ed by Duke et al. (1972); Tr i f unac (1976); Arnol d et al.
(1976).
318 D. M. BOORE, W. B. JOYNER, A. A. OLIVER, AND R. A. PAGE
PUBLISHED CURVES FOR PEAK ACCELERATION
A n u mb e r of publ i s he d cor r el at i ons be t we e n gr ound mot i on p a r a me t e r s a nd
di st ance, ma gni t ude , a nd si t e condi t i ons ha ve be e n des cr i bed b y Tr i f una c a nd Br a d y
(1976) and di scussed by Seed et aL (1976). Th r e e r e c e nt l y publ i s hed, wi del y known
r el at i ons pr opos e d f or p e a k accel er at i on ( Donovan, 1973; Sc hna be l a nd Seed, 1973;
Tr i f unac, 1976) ar e cons i der ed her e.
Cur ve s f or me a n p e a k accel er at i on ar e s hown in Fi gur e 7 f or e a r t h q u a k e s wi t h
1 . 0
b ~
Z
z"

i---
-.J 0.1
U
L)

n
0 . 0 1
" 1 -
0 , 0 0 1 f
1
MAGNI TUDE 6. 4
I I i l i l l i l I I I ~ I
O . . . . . . . " ~ , ' ,
. . . . . . . . . . . .
1 , 0
Z
z"

~ 0. 1
U
U
Z
o
N
O.Ol
-r
MAGNITUDE 7. 2
- - F ~ ~ I L I l l 1 I
" 1 0 - ~ . .
T P__~ ..,,,~
S~ " \ , " ' x
. . . . . . ~ . . ~ . .
" i
I I E P I I I I I P I I I h i l l ] I I 0 . 0 0 1 I I I I L ~ J _ _ ~ i I l l l J I I
1 0 1 0 0 1 0 1 0 0
DISTANCE, IN KtLOMETERS DISTANCE, IN KILOMETERS
FIG. 7. Proposed relations of peak horizontal acceleration to distance from slipped fault for magnitude
6.4 and 7.2 earthquakes. Curve S is given by Schnabel and Seed (1973} for rock sites; curve D by Donovan
(1973) for soil sites; curves TO and T2, mean curves given by Trifunac (1976) for soft and hard sites,
respectively. Solid lines show 70 per cent prediction interval for data set from small structures for
magnitude classes 6.0 to 6.4 and 7.1 to 7.6. The abscissa represents closest distance to the fault for the 70
per cent prediction interval and curve S, epicentral distance for curves TO and T2, and hypocentral
distance for curve D; these differences in distance definitions must be kept in mind when comparing the
curves.
l ocal ma gni t ude s cl ose t o 6.4 a nd 7.2. Al so s hown ar e t he 70 per cent pr edi ct i on
i nt er val s f or t he da t a s et f or ma gni t ude cl asses 6.0 t o 6.4 a nd 7.1 t o 7.6 usi ng da t a
f r om s mal l s t r uc t ur e s ( dat a f r o m t he ba s e of l ar ge s t r uc t ur e s wer e not excl uded i n
t he d e v e l o p me n t of t he ot he r cur ves) .
Th e da t a f r o m t he 1971 Sa n Fe r n a n d o (ML = 6.4) a nd t he 1952 Ke r n Count y (ML
= 7.2) e a r t h q u a k e s e xe r t e d a d o mi n a n t r ol e in t he anal ys es l eadi ng t o t he cur ves
s hown i n Fi gur e 7, a n d t h a t is why t he f i gur es h a v e be e n l abel ed wi t h t hei r
ma gni t ude s . Th e va r i ous wor ke r s ci t ed di d not neces s ar i l y us e t hes e ma gni t ude s f or
t he 1952 a nd 1971 event s; f or cons i s t ency i n pr e pa r i ng t he fi gure, t he ma g n i t u d e s
t h e y di d us e wer e s ubs t i t ut e d i nt o t he e qua t i ons des cr i bi ng t hei r cur ves. Th e l ar ges t
di s c r e pa nc y in t he ma gni t ude was f or t he Ke r n Co u n t y event . Th e a ut hor s of t he
t hr e e a t t e n u a t i o n cur ves wi t h whi ch t he r es ul t s of t hi s s t udy ar e bei ng c ompa r e d,
ESTIMATING PEAK ACCELERATION, VELOCITY, AND DISPLACEMENT 319
used a surface-wave magnitude of 7.6 or 7.7; it was not until 1978 t hat the Richter
local magnitude of 7.2 was published (Bolt, 1978; Kanamori and Jennings, 1978).
This is a good example of the problems t hat can arise when magnitude is used as an
independent variable in the specification of peak motions. It should be noted t hat
the commonly used magnitudes will saturate as the size of the earthquake increases.
A number of recent papers have discussed this important point, e.g., Brune (1970);
Geller (1976); Kanamori (1977); Hanks and Kanamori (1979).
The curve labeled "S" was developed by Schnabel and Seed (1973) for rock sites
and is based on strong-motion data extended to distances nearer the fault with the
help of theoretical attenuation curves. Because the theoretical curves are based on
the conservation of radiated energy, however, t hey apply strictly to quantities based
on an integral measure of the ground motion over the duration of the seismic record
rather than an isolated peak value. Application of the curves to peak parameters is
an approximation of uncertain accuracy. The measure of distance used by Schnabel
and Seed is the shortest distance to the rupture surface, the measure used in this
paper.
The curves labeled "TO" and "T2" are the mean curves given by Trifunac (1976)
for soft and hard sites, respectively. These curves are based on a data set very
similar to the one used in this report, including data from both large and small
structures. The distance measure used by Trifunac is epicentral distance. His curves
were fitted to the data on the assumption t hat the distance dependence of peak
acceleration is t hat of the function given by Richter (1958) for the decay with
distance of the maximum motion of a Wood-Anderson seismograph (free period =
0.8 sec, damping = 0.8). In the absence of more data close to large earthquakes, the
accuracy of Trifunac' s assumption is difficult to evaluate. Even if valid, however, it
should be noted t hat Richter' s attenuation function was not well constrained by
data for distances between 0 and 20 km, which is the range most critical for strong-
motion predictions. Kanamori and Jennings' analysis of local magnitudes from a
few close-in strong-motion recordings (Kanamori and Jennings, 1978) appears to
support the use of Richter' s attenuation function, but small residuals in estimated
magnitudes can correspond to large factors of uncertainty in peak motions, and
Kanamori and Jennings interpreted the distance in Richter' s attenuation relation to
be distance to the surface projection of the center of faulting rather t han epicentral
distance.
The curve labeled "D", developed by Donovan (1973) for soil sites, was obtained
by fitting 678 data points by a function of the form
y = bleb2m(R + 25)-b~
where y is peak acceleration; m, magnitude; R, hypocentral distance in kilometers;
and bl, b2, and b3, adjustable constants. The arbitrary constant 25 is added to the
distance for the purpose of reducing the predicted values at small distances. The
size of the constant has a very large influence on the values at small distances, but
data points at these distances are not sufficient to determine the appropriate size.
Donovan states t hat the function fits the data better when the arbitrary constant is
25 t han when it is zero, but why it should be 25 rather t han any other number is
unclear.
The amount of disagreement shown in Figure 7 is not surprising in view of the
different assumptions, different measures of distance, and different data sets used in
320 D. M. BOORE, W. B. JOYNER, A. A. OLIVER, AND R. A. PAGE
arriving at the different curves. As might be expected, the disagreement is the
greatest at short distances.
CONCLUSIONS
The regression lines given here for peak horizontal acceleration, velocity, and
displacement from the data of earthquakes of western Nort h America provide the
means for estimating peak ground-motion parameters at distances greater than
about 5 km for magnitude 5.0 to 5.7 earthquakes, 15 km for magnitude 6.0 to 6.4
earthquakes, and 40 krn for magnitude 7.1 to 7.6 earthquakes. Unfortunately, most
damage can be expected to occur at shorter distances. As shown, at t empt s have
been made to provide curves for estimating peak accelerations at such distances.
We do not have complete confidence in those curves. The main goal of this paper
has been a careful analysis of existing strong-motion data; the difficult problem of
predicting the ground motion at short distances is being attacked from a number of
angles by us and by other workers. The synthesis of this work into attenuation laws
and the subsequent confirmation of these laws lies in the future.
ACKNOWLEDGMENTS
We are grateful to R, P. Maley for assistance in obtaining information on strong-motion recording site
conditions and to A. G. Brady for unpublished strong-motion data from the February 21, 1973, Point
Mugu, California, earthquake. R. B. Matthiesen and B. A. Bolt made useful suggestions on earlier
versions of the paper. This work was done in cooperation with the Nuclear Regulatory Commission.
REFERENCES
Acton, F. S. (1959). Analysis of Straight-Line Data, Dover, New York, 267 pp.
Arnold, P., E. H. Vanmarcke, and G. Gazetas {1976}. Frequency content of ground motions during the
1971 San Fernando earthquake: Massachusetts Institute of Technology, Department of Civil Engi-
neering, Publication R76-3, 73 pp.
Bolt, B. A. {1978). The local magnitude ML of the Kern County earthquake of July 21, 1952, Bull. Seism.
Soc. Am. 68, 513-515.
Boore, D, M. and M. D. Zoback {1974). Two-dimensional kinematic fault modeling of the Pacoima Dam
strong-motion recording of the February 9, 1971, San Fernando earthquake, Bull. Seism. Soc. Am.
64, 555-570.
Boore, D. M., W. B. Joyner, A. A. Oliver, III, and R. A. Page (1978). Estimation of ground motion
parameters, U.S. Geol. Surv., Circular 795, 43 pp.
Brune, J. N. (1970). Tectonic stress and the spectra of seismic shear waves from earthquakes, J. Geophys.
Res. 75, 4997-5009.
Crouse, C. B. (1978). Prediction of free-field earthquake ground motions, Proc. ASCE Geotech. Eng. Div.
Specialty Conf. Earthquake Eng. Soil Dynamics. Pasadena, Ca., 1, 359-379.
Crouse, C. B. and P. C. Jennings (1975). Soil-structure interaction during the San Fernando earthquake:
Bull. Seism. Soc. Am. 65, 13-36.
Dixon, W. J. and F. J. Massey, Jr. (1957). Introduction of Statistical Analyses, McGraw-Hill, New York,
488 pp.
Donovan, N. C. (1973). A statistical evaluation of strong motion data including the February 9, 1971 San
Fernando earthquake, Proc. Worm Conf. Earthquake Eng., 5th, Rome, 1, 1252-1261.
Duke, C. M., J. E. Luco, A. R. Carrivean, P. J. Hradilek, R. Lastrico, and D. Ostrom (1970). Strong
earthquake motion and site conditions, Hollywood, Bull. Seism. Soc. Am. 60, 1271-1289.
Duke, C. M., K. E. Johnsen, L. E. Larson, and D. C. Engman (1972). Effects of site classification and
distance on instrumental indices in the San Fernando earthquake: University of California. Los
Angeles, School of Engineering and Applied Science, UCLA-ENG-7247, 29 pp.
Geller, R. J. (1976). Scaling relations for earthquake source parameters and magnitudes, Bull. Seism.
Soc. Am. 66, 1501-1523.
Hanks, T. C. (1974). The faulting mechanism of the San Fernando earthquake, J. Geophys. Res. 79,
1215-1229.
Hanks, T. C. {1975). Strong ground motion of the San Fernando, California, earthquake, ground
displacements, Bull. Seism. Soc. Am. 65, 193-225.
ESTIMATING PEAK ACCELERATION, VELOCITY, AND DISPLACEMENT 321
Hanks, T. C. and H. Kanamori (1979). A moment magnitude scale, J. Geophys. Res. 84, 2348-2350.
Kanarnori, H. (1977). The energy release in great earthquakes, J. Geophys. Res. 82, 2981-2987.
Kanamori, H. and P. C. Jennings (1978). Determination of local magnitude, ML, from strong-motion
accelerograms, Bull. Seism. Soc. Am. 68, 471-485.
Newmark, N. M. and W. J. Hall (1969). Seismic design criteria for nuclear reactor facilities, Proc. World
Conf. Earthquake Eng., 4th, Santiago, 2, B4-37-B4-50.
Newmark, N. M., J. A. Blume, and K. K. Kapur (1973). Seismic design spectra for nuclear power plants,
Proc. Am. Soc. Civil Eng. J. Power Div. 99, 287-303.
Page, R. A., D. M. Boore, W. B. Joyner, and H. W. Coulter (1972). Ground motion values for use in the
seismic design of the trans-Alaska pipeline system, U.S. Geol. Surv., Circular 672, 23 pp.
Page, R. A., D. M. Boore, and J. H. Dieterich (1975). Estimation of bedrock motion at the ground surface,
in Studies for Seismic Zonation of the San Francisco Bay Region, R. D. Borcherdt, Editor, U.S.
Ge)l. Surv., Profess. Paper 941-A, A31-A38.
Richter, C. F. (1958). Elementary Seismology, W. H. Freeman, San Francisco. 768 pp.
Schnabel, P. B. and H. B. Seed, (1973). Accelerations in rock for earthquakes in the western United
States, Bull. Seism. Soc. Am. 63, 501-516.
Seed, H. B., R. Murarka, J. Lysmer, and I. M. Idriss {1976). Relationships of maximum acceleration,
maximum velocity, distance from source, and local site conditions for moderately strong earthquakes,
Bull. Seism. Soc. Am. 66, 1323-1342.
Trifunac, M. D. (1976). Preliminary analysis of the peaks of strong earthquake ground motion--
dependence of peaks on earthquake magnitude, epicentral distance, and recording site conditions,
Bull. Seism. Soc. Am. 66, 189-219.
Trifunac, M. D. and J. N. Brune (1970), Complexity of energy release during the Imperial Valley,
California, earthquake of 1940, Bull. Seism. Soc. Am. 60, 137-160.
Trifunac, M. D. and A. G. Brady (1976). Correlations of peak acceleration, velocity, and displacement
with earthquake magnitude, distance, and site conditions, Earthquake Eng. Structural Dyn. 4, 455-
471.
U.S. GEOLOGICAL SURVEY
345 ]~_.IDDLEFIELD ROAD
MENLO PARK, CALIFORNIA 94025
Manuscript received May 14, 1979
NOTE ADDED IN PROOF
Several recent eart hquakes in t he west ern Uni t ed St at es have provi ded dat a
whi ch can be used to t est our conclusions concerni ng t he prediction of peak mot i ons
in vari ous distance ranges for different magni t ude classes. In a not e in preparat i on,
Boore and Porcel l a have compar ed t he peak accelerations from t he 1978 Sant a
Bar bar a (ML ~- 5.1), 1979 Coyot e Lake (ML = 5.9), and 1979 Imperi al Valley (ML
= 6.4) eart hquakes wi t h t he regression curves in Figure 1 of this paper. They find
t hat in t he distance range over whi ch t he regression lines are defined t he agreement
in each case is excellent. For t he combi ned Coyot e Lake and Sant a Bar bar a
ear t hquake data, 9 out of 10 values are within t he 70 per cent predi ct i on intervals,
and 12 out of 17 poi nt s (71 per cent) from t he Imperi al Valley eart hquake fall within
t he correspondi ng 70 per cent predi ct i on intervals and all 17 points fall within t he
95 per cent predi ct i on intervals. The Imperi al Valley eart hquake provi ded 19 dat a
poi nt s at distances less t han 15 km (compared to only 2 values for t he analysis in
this paper), and t hus will help extend to closer distances our ability to predict
ground motions. Boore and Porcel l a find t hat t he close-in points from t he Imperi al
Valley ear t hquake fall below a linear ext rapol at i on of t he mean regression line for
t he 15 to 55 km distance range.