Bulletin of the SeismologicalSociety of America. V01.57, No. 6, pp. 1193-1220. December~1967
ANALYSIS
OF
ACCELEROGRAMS--PARKFIELD
BY G. W. HOUSNER
EARTHQUAKE
AND M. D. TmFUNAC
ABSTRACT
Integrated velocities and displacements show that near the fault at Cholame
the surface motion exhibited a transient horizontal displacement pulse of approximately ten inches amplitude and one and one-half seconds duration, normal
lo the fault. Although 50 per cent of g ground acceleration was recorded at the
fault, Ihe ground motion attenuated rapidly with distance and at ten miles from
the fault the maximum acceleration was reduced ~o one-tenth of its near-fault
value. The ground motion also changed its character with distance, losing its
pulse-like directional characteristic and becoming isotropic. Computed response
spec|ra are presented and the large spectrum ordinates for this shock of relatively small magnitude and moderate destructiveness indicate that in an engineering sense the Parkfield ground motion is in a different class from such large destructive ground motions as El Centro 1940, Tehachapl 1952, and Olympia 1949.
INTRODUCTION
The ground motion of the magnitude 5.6 Parkfie.ld earthquake of June 27, 1966
was recorded by an array of five strong-motion acce]erographs and fifteen seismoscopes across the San Andreas fault at Cholame. Hence, the instrumental coverage
of this earthquake was much more extensive than that of any other. The maximum
recorded ground acceleration of the fault was 50 per cent of gravity, which
is stronger than has been recorded for any other earthquake. On the other hand, the
ground motion caused little damage to structures, presumably because of its very
short duration. The impulsive strong motion at the fault near Cholame had a large
component transverse to the fault which consisted essentially of a single northeasterly displacement pulse of 10 inches amplitude and 1.5 seconds duration. The
instrument at the fault near Cholame was near the end of the surface cracking which
extended some 20 miles northwesterly along the San Andreas fault beyond Parkfield. The original accelerograph recordings (Cloud and Perez, 1967), the seismoscope recordings (Hudson and Cloud, 1967), and the damage caused by the earthquake (Cloud, 1967) are described in companion papers.
The array, shown in Figure 1, had been installed only shortly before the earthquake as a cooperative venture of the California Department of Water Resources
and the U. S. Coast and Geodetic Survey. The D W R project to bring water from
the Feather River to southern California involves the construction of facilities
close to known active faults, in particular, the San Andreas fault, which is crossed
by the aqueduct several times. The design of these facilities, as well as the planning
of nuclear reactor power plants in California, raised questions as to the nature of the
ground motion immediately adjacent to the causative fault and how it changes with
distance out to approximately ten miles from the fault. It was recommended by one
of the authors (GWH) that arrays of accelerographs be placed across the San
1193
1194
BULLETIN
OF
THE
SEISMOLOGICAL
SOCIETY
OF AMERICA
Andreas fault at several locations, with o n e array across the fault somewhere between the southerly end of the fault slip of the 1906 San Francisco earthquake and
the northerly end of the fault slip of the 1857 Fort Teion earthquake. The most
\,,,%
\
\dL
%
x
: ~ PARKFI:LD
N. . . .
%,
ShO~.ANIE S--L}NG%1L)T'CLI ARFAY
%%\
• Accelerograph
o Seismcscope
\
", . # ~ =AULT TRACE
\
X\
ANDREAS
FAULT ZONE
X
j
STATIONS
3 EMBLOR ¢
\
o
\
o 5
6
\
o
¢
\
7
9
o:O
II
e
2
\
\
0
i
2
3
MILES
4
5
T
I
\
"\
\
FIG. 1. Strong-Motion Array at Cholame. Five accelerographs and fifteen seismoscopes.
accessible line was along the highway through Cholame. The northerly end of the
1857 slip is thought to have been not far distant. A second array has been inst~lled
across the San Andreas fault in the Lake Hughes area just north of Los Angeles.
The aecelerographs are UED AR-240's which record one vertical ~nd two hori-
ANALYSIS
OF
ACCELEROGRAMS--PARKFIELD
1195
EARTHQUAKE
zontal components of motion. The instruments have a natural frequency of approximately 18 cps, and have approximately 65 per cent of critical damping. These instrumerits record essentially the true ground acceleration. Unfortunately, there was a
malfunction of the instrument at Station 2 so that the component parallel to the
fault did not record.
The epicenter of the magnitude 5.6 Parkfield shock was located on the San Andreas fault approximately 20 miles northwest of Cholame. This shock was preceded
-o+
z
•
I
I
I
I
I
i
I
<+ 0.2
m
N
~~I
o
~
~ -0.2
0
n.-
°o[
P+
2O
0
t'r'
ILl
O
3
-J
a.
fD
0
t",
-3
j
Z,-~
0
I
I
I
I
I
1
I
I
0
2
4
6
8
I0
12
14
T I M E , SECONDS
F I G . 2. T e m b l o r
Station
Vertical
Motion.
by a magnitude 5.1 earthquake whose epicenter had approximately the samelocation.
The earlier shock triggered the accelerographs at Stations 2, 5 and 8 which recorded
maximum horizontal accelerations of slightly less than 2 per cent of g, which is
about triggering sensitivity. Accelerographs at Stations 12 and Temblor were not
triggered by the smaller shock.
There was surface cracking along the fault from about the epicenter to just south
of Station 2. Hence, a slipped length of fault of some 20 miles was indicated.
1196
BULLETIN
OF THE
SEISMOLOGICAL
SOCIETY
OF AMERICA
THE RECORDED GROUNI) 2\'~OT[ONS
As shown in Figure 1, there is an offset in the fault trace of about one mile between
Cholame and Parkfield. The valley between Cholame and Parkfie]d appears to be a
i
i
i
I
i
i
!
"o, 0.4
<~ 0.2
_J
uJ
0
0
'¢¢"~"
-0.2
<.~ -0.4
40
.=_
20
>.
I-0
.
IJJ
O
~
~-ao
(5
~-40
O
z
LO 5
hi
O
5fl_o
5q
-5
z
0.6
(.9
k
0
I
I
I
I
I
1
4
6
8
I0
12
14
T I M E , SECONDS
F I G . 3. T e m b l o r
Station
$25W Motion.
fault zone of about a mile in width. It is not known what influence, if any, this offset
had upon the recorded ground motions.
The analysis of the ground motions was done at Calteeh using digitized accelerations provided by the Seismographic Field Station of the USC&GS. As a cheek,
several of the aecelerograms were independently digitized at Calteeh and compared
ANALYSIS
OF
ACCELEROGRAMS--PARKFIELD
1197
EARTHQUAKE
with the USC&GS digitized data and good agreement was found. A fifteen-second
long section of each accelerogram was selected for analysis. A cubic-curve baseline
was fitted to each ground motion by a least squares fit on the velocity (ttousner and
I
I
I
I
~
I
....
I
¢ft
"o~ 0.4 I
g
i
: 0.2
a-"
LIJ
W
0
0
C~
c~ -0.2
Z
0
Pr" -0.4
(.9
o 40
¢D
.E
>_ 2 0
p-
3bJ o
-20
O
-40
,a 6
Fz
ua 3
i,a
(9
5o
~9
-3
Z
~
-6-
I
I
I
I
I
{
I
2
4
6
8
I0
12
14
TIME, SECONDS
FI~. 4. Temblor Station N65W Motion.
Berg, 1961). The resulting acceleration, velocity and displacement represented by
the processed digitized data is shown for each recorded component of each accelerograph in Figures 2-15, as drawn by the computer.
It is seen in Figure 6 that at Station 2 the transverse component had a maximum
acceleration of 50 per cent of g. Figures 3 and 8 show that the maximum acceleration
1198
BULLETIN
OF
THE
SEISMOLOGICAL
SOCIETY
OF
AMERICA
exceeded 40 per cent of g out to a distance of about 3.5 miles each side of the fault.
At Station 8 the maximum acceleration was 27 per cent of g, and at. Station 12 it
was reduced to 5 per cent of g. This rapid attenuation with distance from the fault
indicates a relatively shallow depth of slipped fault as the origin of this motion.
I
"~
I
I
I
I
I
[
k
0.4
Z
0
0.2
!aJ
_J
uJ
o
0
c~ -0.2
z
ZD
0
rr -0.4
c_o
.c_
20
I
r
c
6
l-z
5n 0
o)
-3
Z
-6--
2
I
I
I
I
I
I
4
6
8
IO
12
14
TIME, SECONDS
F r o . 5. S t a t i o n 2 V e r t i c a l M o t i o n .
Figure 6 shows the maximum ground velocity at Station 2 to be 27 inches per
second. It also shows the displacement to be essentially a single pulse of approximately 10 inches amplitude and 1.5 seconds duration. This is followed by motion
having about an inch or less amplitude and about one second predon'dnant period.
Figures 8 and 9 show that at Station 5 the horizontal ground acceleration has
ANALYSIS
OF ACCELEROGRANIS--PARKFIELD
1199
EARTHQUAKE
already become essentially isotropic and the ground displacement does not appear
as a single pulse, but has a shape similar to that of more orthodox ground motions
such as Taft, 21 July 1952 and Olympia, 13 April 1949. At Stations 8 and 12 the
ground displacement curves become more smooth with increasing distance from the
fault.
J
1
I
I
I
I
I
I
o.61--~
0.4-
z
0
0.2-rY
bJ
_J
bJ
(D
0 :~
I
°e-~ -0,2 i0
n- -0,4
-0.6
~ 4O
" ; 20
0
-40
c
6
bZ
t~ S
ILl
;o
Q.
{D
l:l
r~
Z
-3
-9
-12 -
I
I
I
I
I
I
I
I
0
2
4
6
8
I0
12
14
TIME, SECONDS
FIG. 6. Station 2 N65E Motion.
19~00
BULLETIN
OF THE
SEISMOLOGICAL
SOCIETY
OF AMERICA
Figures 3 and 4 show that at the Temblor Station the motions were similar to,
but somewhat less strong than those at Station 5. On the Temblor side of the fault
basement granite at a depth of 2.6 miles is overlain by four sedimentary layers and
I
I
I
~
I
I
I
0.4
"~
Z
0
Q
0.2
n,u..I
_.J
LU
(D
0
C3 -0.2
Z
0
{.9
-0,4
g 40 r.E
]
>. 20
So
uJ
>
~-4o~
W 3-UJ
m
N
z
0
I
1
I
I
I
I
I
2
4
6
8
IO
12
14
TIME, SECONDS
FIo. 7. Station 5 Vertical Motion.
on the southwestern side granite at 2.2 miles depth is overlain by three sedimentary
layers (M. E. O'Neill, private communication) as shown in Table 1. In addition
there is a shallow surface layer of alluvium at Stations 5, 8 and 12. Station 2 is
directly on the fault zone.
The vertical acceleration at Station 2, shown in Figure 5, has a maximum value
ANALYSIS
OF ACCELEROGRAMS--PARKFIELD
1201
EARTHQUAKE
of 35 per cent of g. The appearance of the accelerogram differs markedly from that
of the horizontal motion, Figure 6. There is a three-second long segment of relatively
high frequency motion whose amplitude becomes very small about five seconds after
I
I
I
I
I
I
I
I
I
- ~ 0.4
Z
O
z~
<~ 0.2
n"
IJJ
J
,,,
0
U
<~
-
0
C~ - 0 . 2
Z
-
0
n," - 0 . 4
(.9
-
o 4O
.c_
>.
#-
20
0
_j
bJ
0
~-ao
O
-40
3 itlJ
0
<,0
(1.
{D
-3 Z.-~
0-61.9
0
I
I
I
I
f
I
I
2
4.
6
8
I0
12
14
TIME, SECONDS
FIG. 8. Station 5 N85E Motion.
the start of the record. There is also visible a longer period component, approximately one second period, that has about 5 per cent of g amplitude out to six seconds after the beginning of the accelerogram, and then gradually dies away. This
longer period component shows clearly in the vertical ground displacement. The
1202
: B U L L E T I N OF T H E S E I S M O L O G I C A L S O C I E T Y OF A M E R I C A
vertical ground motion attenuates with distance from the fault even more rapidly
than does the horizontal motion. At Stations 8 and 12 the vertical motion is very
much attenuated and even at Station 5 and at Temblor Station, only about 3.5 miles
from Station 2, the vertical motion is much reduced.
I
I
l
I
I
I
I
I
I
I
I
6
8
I0
12
14
0.4
~
Z
0
0.2
nLU
..J
UJ
0
¢.,)
0
,,~ -0.2
Z
:D
0
n~ -0.4
~9
o 4O
,E
).- 2 0
L)
30
oA
¢.}
2
0
n
r
J
J
~
(D
-3
Z
0-6-
0
I
I
2
4
,I
T I M E , SECONDS
F~G. 9. S t a t i o n 5 N 5 W M o t i o n .
RESPONSE SPECTRA
The computed response spectra are presented in Figures 16-29. In each figure the
four curves are the velocity response spectra for 0, 2, 5 and 10 per cent of critical
damping. It may be noted that the zero damped response spectrum curve is a fair
representation of the Fourier amplitude spectrum of the accelerogram.
ANALYSIS OF A C C E L E R O G R A M S - - P A R K F I E L D EARTHQUAKE
1203
Figure 20 shows the spectrum of the transverse motion at Station 2. The spectrum
has a pronounced peak at about 0.75 seconds period which is produced by the few
large peaks on the accelerogram. There is a lesser peak at 1.25 seconds period and a
I
I
I
I
I
I
l
I
I
"~ 0.4
0.2
(.)
o
a -0.2
z.-j
o
n- -0.4
(.9
o 40.c_
>.. 2 0 o
w
>
~-20-0
-40
._e 6
I
I
I
I
I
I
I
2
4
6
8
I0
12
14
TIME, SECONDS
FIG. 10. S t a t i o n 8 V e r t i c a l M o t i o n .
broad hump at about 1.5 seconds period. It should be noted also that the spectrum
curves drop off rapidly at the short period end, being already very low at a period of
0.2 seconds. The spectra at Station 5 show that the horizontal components have
higher frequencies than at Station 2 for there is a pronounced peak at about 0.4 seconds period. In the N 85 E spectrum at Station 5 there are two pronounced peaks, at
1204
BULLETIN
OF T H E S E I S M O L O G I C A L S O C I E T Y OF A M E R I C A
1.25 and 2.0 seconds respectively, that do not appear on the NSW component. These
velocity speetra at Station 5 with sharp peaks at 0.4 seconds period are not ~oo
dissimilar from those calculated for other small earthquakes recorded relatively
I
u>
o~ 0.4
I
I
I
I
I
I
I
i
Z
O
0,2
n.,"
tl.I
.3
W
0
<I
0
r~ -0.2
Z
Z)
0
r r -0.4
<D
4O
.¢_
~_ 2O
t-
o
O,
tad
>
2,-a~
=o
O -40
c sl-
H
t~
oI
r~
Z
~-6-(D
I
I
I
I
I
I
I
2
4
6
8
I0
12
14
TIME, SECONDS
FIG. 11. Station 8 N40W Motion.
close to epicenters as, for example, the Golden Gate Park record for the March 22,
1957 San Francisco shock (M = 5.3).
The spectrum curves for Station 8 do not exhibit the marked peak at 0.4 seconds
period that the Station 5 spectra showed. This difference in shape of spectra is
presumably due to the greater distance from the fault of Station 8. So far as is
1205
ANALYSIS OF ACCELEROGRAMS--PARKFIELD EARTHQUAKE
known, there is no significant difference in geology at the two stations which are
less than three miles apart.
The spectrum curves for horizontal motion at Station 12 show some pronounced
I
I
I
I
I
1
'-
I
_m
o, 0.4
Z
0
0.2
W
,.
0
0
0
£3 -0.2
Z
-0.4
~ o
~
-40
c 6
z
-3
Z
:::)
I
I
I
I
I
I
I
2
4
6
8
I0
12
14
TIME, SECONDS
FzG. 12. S t a t i o n 8 N50E M o t i o n .
humps at 1.4 and 2.2 seconds period that did not appear at Station 8, although
similar humps appeared in the N85E component at Station 5.
The Temblor horizontal spectra are similar to those of Station 5 but the large
hump at 2 seeonde in ths N85E component at Station 5 does not appear at Temblor.
The vertical component spectrum for Station 2 shows peaks at 0.75 and 1.5
1206
BULLETIN
OF
THE
SEISMOLOGICAL
SOCIETY
OF
AMERICA
seconds similar to the horizontal component spectrum. The vertical components
at Stations 5, 8, 12 and TembIor do not show any pronounced frequency components.
I
i
I
I
I
I
J
j
0.4
Z
O
,~ 0.2
/lg
tJA
_A
,,,
0
o
o -0.2
Z
0
¢: -0.4
o 4O
c
>. 20
p-
o,
o
w
>
~-zo
O
er
¢9 - 4 0
.=2
z
3
t~J
C)
<,o
o_
G9
a
-3
Q
z
(.9
I
I
2
4
I
!
6
I
8
i
I0
12
l
14
TIME, SECONbS
FI(~. ]3. S t a t i o n ]2 V e r t i c a l M o t i o n .
THE
iA~rATURE O F
THE
~'IOTION
AT
STATION
2
Station 2 was located about 200 feet westerly of the line of fault slip that passed
through the asphalt road near Cholame. The recorded motion transverse to the
fault showed a maximum acceleration of 50 per cent of g corresponding to a displacement pulse of about 10 inches amplitude and 1.5 seconds duration. The
ANALYSIS
ACCELEROGRAMS--PARKFIELDEARTHQUAKE
OF
1207
direction of this pulse was towards the northeast. As the component longitudinal
to the fault was not recorded, it is of interest to see what can be deduced about
this motion. Figure 30 shows the retouched seismoscope record at Station 2. The
I
I
I
I
J
I
I
I
I
I
I
I
t
2
4
6
8
I0
12
14
o~ 0.4
Z
0
0.2
nr"
hi
..J
W
0
(.)
0
Q -0,2
Z
o
iv" - 0 . 4
(_9
~2
z
Lo
uJ
_J
o_
Q
Z
0
TIME, SECONDS
FIG. 14. Station 12 N40W Motion.
solid line portion of the record is reliably recorded, and the dotted line indicates
that the recording needle skipped, went off the edge and then returned onto the
smoked watch-glass, touched the edge again and shortly thereafter the glass was
dislodged and dropped on the floor. The following check was made of the reliability of the seismoseope record: The response of the seismoscope, which had
0.78 seconds period and 8.5 per cent damping, was calculated using the recorded
1208
BULLETIN
OF THE
SEISMOLOGICAL
SOCIETY
OF AMERICA
transverse component of acceleration as input. This computed response agreed
very closely with the projection of the seismoscope record on a plane transverse
to the fault, including both the solid and the dotted lines shown in Figure 30.
I
I
I'
I
I
I
1
0.4
Z
0
Q
O.Z
LU
.J
L,J
(.3
0
-0.2
z
0
n-" -0.4
40--
o
.c
>. 2 0 ~9
o
ul
>
~-20-O
z
w
3
c~
J
Q.
z
0_4
O
I
I
I
I
I
I
I
2
4
6
8
I0
12
14
TIME, SECONDS
F I G . 15. S t a t i o n
12 N S O E M o t i o n .
This covers the initial portion of the accelerogram up to the 50 per cent g pip. The
seismoseope record thus shows that during the initial portion of the ground motion the transverse component was much stronger than the longitudinal component.
Similar indications have been deduced from seismoseope records from other shocks
(Cloud, 1967b).
Figure 31 shows the seismoscope record at Station 1 which was about one mile
easterly of the fault. Although the sequence of motion cannot be deciphered, it is
ANALYSIS OF ACCELEROGRAMS--PARKFIELD EARTHQUAKE
1209
clear that the transverse motion was much stronger than the longitudinal component of motion. Figure 32 shows the seismoscope record for Station 3 about one
mile westerly of the fault. Recalling that the dotted lines in Figure 30 take the
record just up to the 50 per cent of g pulse, it may be concluded that the effective
TABLE 1
GENERALIZED P-WAvE VELOCITY STRUCTURE
Velocity (It per sec)
Thickness of Layer (ft)
A. Southwest of San Andreas Fault Zone Near
Cholame
5,400
8,900
16 000
19,000
900
4000
7000
B. Norlheast of San Andreas Fault Zone Near
Cholame
7,500
10,500
14,500
18,000
19,000
570
3300
4900
5300
60
PARKFIELD EARTHQUAKE, JUNE 1966
CHOLAME SHANDON ARRAY
STATION NO. TEMBLOR VERT
5O
Sd'-~, RELATIVE RESPONSESPECTRUM
d
SII = DISPLACEMENT SPECTRUM
.F. 4 0
¢/)
IS
w
3o
w
z
~ 2o
w
I0
0
I
2
5
4
PERIOD T, SECONDS
FIG. 16. Response Spectra, Temblor-Vertical. The curves are for O, 2, 5 and 10% damping.
1210
BULLETIN OF THE SEISMOLOGICAL SOCIETY OF AMERICA
60
PARKFIELD EARTHQUAKE, JUNE 1966
CHOLAME SHANDON ARRAY
STATION NO. TEMBLOR SZ5W
50
Sd-~-~ , RELATIVE RESPONSE SPECTRUM
o)
Sd = DISPLACEMENT SPECTRUM
.E 40
>.-
q
30
i,i
u.I
m
Z
~ ~o
ILl
IO
0
I
I
I
I
I
2
3
4
PERIOD T, SECONDS
FIG. 17. Re@onse Spectra, Temblor-S25W. The curves are for 0, 2, 5 and 10% damping.
60
PARKFIELD EARTHQUAKE, JUNE 1966
CHOLAME SHANDON ARRAY
STATION NO. TEMBLOR N65W
50
Sd"~,
RELATIVE RESPONSE SPECTRUM
Sd : DISPLACEMENT SPECTRUM
.c 40
>
~0
>F-
q
30
W
W
Z
o
20
03
LIJ
rr
I0
I
2
3
4
PERIOD T, SECONDS
FI~. 18. l~esponse Spectra, Temblor-N65W. The curves are for 0, 2, 5 and 10% damping.
ANALYSIS
OF
ACCELEROGRAMS--PARKFIELD
1211
EARTHQUAKE
I00
PARKFIELD EARTHQUAKE, JUNE 1966
CHOLAME SHANDON ARRAY
STATION NO. 2 DOWN
80
Sd"-~,
RELATIVE RESPONSE SPECTRUM
Sd = DISPLACEMENT SPECTRUM
c:
O9
>- 6 0 - p-
g
LIJ
>
uJ 4 O - co
Z
O
0_
o9
uJ
n20--
0
I
I
I
I
I
2
3
4
PERIOD T, SECONDS
FIG. 19. Response Spectra, Station 2-Vertical. The curves are for 0, 2, 5 ~nd ]0% damping.
iO0
PARKFIELD EARTHQUAKE, JUNE 1966
CHOLAME SHANDON ARRAY
STATION NO. 2 N 6 5 E
l/
80
Sd'~,
RELATIVE RESPONSE SPECTRUM
.=_
;P
O9
>- 6 0
I-
q
uJ
>
.N 40
Z
0
u.I
20
0
I
I
i
2
I
3
I
4
PERIOD 1", SECONDS
FIG. 20. Response Spectra, Station 2-N65E. The curves are for 0, 2, 5 and 10% damping.
1212
BULLETIN
OF
THE
SEISMOLOGICAL
SOCIETY
OF
AMERICA
GO
PARKFIELD EARTHQUAKE, .JUNE 1966
CHOLAME SHANDON ARRAY
STATION NO.5 DOWN
50
Sd'~,
if)
RELATIVE RESPONSE SPECTRUM
Sd = DISPLACEMENT SPECTRUM
40
>-
q
30
LU
>
W
03
Z
20
¢./)
W
re"
I
I
I
I
2
3
P E R I O D T, SECONDS
FIG. 21. Response Spectra, Station 5-Vertical. The curves are for 0, 2, 5 and 10% damping.
60
PARKFIELD EARTHQUAKE, JUNE 1966
CHOLAME SHANDON ARRAY
STATION NO.5 N85E
50
Sd'~,
RELATIVE RESPONSE SPECTRUM
Sd = DISPLACEMENT SPECTRUM
c
'-
40
30
ILl
W
Z
o ao
(D
UJ
nr
I0
|
I
1
I
I
0
I
2
5
4
PERIOD T, SECONDS
FIG. 22. Response Spectra, Station 5-N85E. The curves are for 0, 2, 5 and 10% damping.
ANALYSIS OF ACCELEROGRAMS--PARKFIELD EARTHQUAKE
1213
6'[?
PARKFIELD EARTHQUAKE, JUNE 1966
CHOLAME SHANDON ARRAY
STATION NO. 5 NSW
50
Sd-~, RELATIVE RESPONSESPECTRUM
==
Sd = DISPLACEMENT SPECTRUM
.E 4 0
O3
I'-
q 3o
/
w
O3
z
no 2o
O3
W
I
I
I
I
I
2
3
4
PERIOD Z SECONDS
FIG. 23. Response Spectra, Station 5-N5W. The curves are for 0, 2, 5 and 10% damping.
50
PARKFIELD EARTHQUAKE, JUNE 1966
CHOLAME SHANDON ARRAY
STATION NO. 8 DOWN
40-Sd~"~, RELATIVE RESPONSE SPECTRUM
Sd = DISPLACEMENT SPECTRUM
.E
>
O3
>- :30-F-
q
Irl
uJ 2 0 - O3
z
o
0_
(s)
n,,
io--
0
-
I
I
I
I
I
2
3
4
PERIOD T, SECONDS
FZG. 24. Response Spectra, S t a t i o n 8-Vertical. The curves are for 0, 2, 5 and 10% damping.
1214
BULLETIN
OF
THE
SEISMOLOGICAL
SOCIETY
OF
AMERICA
50
PARKFIELD EARTHQUAKE, JUNE 1966
CHOLAME SHANDON ARRAY
STATION NO. 8 N40W
4O
c~
S d - ~ , RELATIVE RESPONSE SPECTRUM
Sd = DISPLACEMENT SPECTRUM
.c
>
>- 3O
q
W
m
m
Z
0
20
r~
I0
0
I
2
3
4
PERIOD T, SECONDS
FIG. 25. Response
Spectra, Station 8-N40W.
The curves are for 0, 2, 5 and 10% damping.
50
PARKFIELD EARTHQUAKE, JUNE 1966
CHOLAME SHANDON ARRAY
STATION NO, 8 N 5 0 E
4O
Sd-~K, RELATIVE RESPONSE SPECTRUM
Sd = DISPLACEMENT SPECTRUM
~0
>- 3 O
b-
sW
W 20
U)
Z
0
OU.I
I0
jr
0
j
~
I
I
I
I
I
Z
3
4
PERIOD T, SECONDS
FIG. 26. Response
Spectra, Station 8-NSOE. The curves are for 0, 2, 5 and 10% damping.
ANALYSIS OF ACCELEROGRAMS--PARKFIELD EARTHQUAKE
1215
30
PARKFIELD EARTHQUAKE, JUNE 1966
CHOLAME SHANDON ARRAY
STATION NO. 12 V E R T .
25
Sd"~,
RELATIVE RESPONSE SPECTRUM
Sd = DISPLACEMENT SPECTRUM
.c_ 20
0")
>..
I-
S
W
15
L~
I0
f./)
ILl
re"
5
i
I
I
i
I
2
3
4
PERIOD T,
SECONDS
FIG. 27. Response Spectra, Station 12-Vertical. The curves are for 0, 2, 5 and 10% damping.
30
PARKFIELD EARTHQUAKE, JUNE 196G
CHOLAME SHANDON ARRAY
STATION NO. 12 N40W
25
S d ' - ~ , RELATIVE RESPONSE SPECTRUM
,5
Sd • DISPLACEMENT SPECTRUM
'-
>
(/)
20
I'-
o, 15
LU
>
LU
Z
0
O.
(/)
hi
I0
I
i
I
i
I
2
3
4
PERIOD T, SECONDS
FIG. 28. Response Spectra, Station 12-N40W. The curves are for 0, 2, 5 and 10% damping.
1216
BULLETIN OF THE SEISMOLOGICAL SOCIETY OF AMERICA
motion at Station 3 was much reduced; that is, the 10 per cent damped spectrum
curve in the neighborhood of 0.75 seconds period was much reduced over the value
at 0.78 seconds at Station 2. The response spectrum displacement values at Station
3 were 0.9 inches single amplitude transverse and 0.65 inches longitudinal, whereas
30
PARKFIELD EARTHQUAKE, JUNE 1966
CHOLAME SHANDON ARRAY
STATION NO. IP N 5 0 E
25
Sd-~,
t/)
c
03
>I0
o,
RELATIVE RESPONSE SPECTRUM
Sd "DISPLACEMENT SPECTRUM
20
15
bd
b.l
Z
10
I
51
0
I
I
I
I
I
2
3
4
PERIOD T, SECONDS
FIG. 29. Response Spectra, S t a t i o n 12-N50E. The curves are for 0, 2, 5 and 10% damping.
\
\
\
"
. :.:.
FIG. 30. Seismoscope Record s t a t i o n 2.
ANALYSIS
OF ACCELEROGRAMS--PARKFIELD
EARTHQUAKE
FIG. 31. Seismoscope Record Station 1.
Z
(/~:::
S/-
'-
•
f
FIG. 32. Seismoscope Record Station 3.
1217
1218
BULLETIN OF THE SEISMOLOGICAL SOCIETY OF AMERICA
the computed response spectrum displacement at Station 2 is 5.4 inches transverse.
A hodograph of the ground displacement at Station 5 is shown in Figure 33. This
shows larger displacements transverse to the fault.
By private communication, Dr. Bruce Bolt reports that at the Berkeley seismographic station, which lies 150 miles northwest of Parkfield near the San Andreas
fault, the longitudinal displacement component had approximately one-third the
amplitude of the transverse component.
0
I
2
5
!
I
|
~
4
[
5
1
INCHES
FIG. 33. Hodograph Station 5 First fifteen seconds of horizontal ground motion,
3.3 miles from fault.
Art attempt was made to squeeze out information from the seismoscope record
at Station 2 together with the known transverse component of acceleration. This
analysis indicated a parallel acceleration pulse of approximately 0.5g lagging
behind the transverse 0.5g pulse about one-quarter of a wave length.
T H E PORT ~TIuENEME EARTHQUAKE
The magnitude 4.7 Port Hueneme earthquake of 1957 was recorded close to the
epicenter and its motion was also a single displacement pulse similar to that at
Station 2 only smaller. The pulse had an amplitude of 1.4 inches, a duration of
0.5 seconds; and its direction was northwesterly, parallel to the San Andreas fault
ANALYSIS OF ACCELEROGRAMS--PARKFIELD EARTHQUAKE
1219
which was some 30 miles to the east. The maximum acceleration was 16 per cent
of g. In a paper (Housner and Hudson, 1958) analyzing this record it was pointed
out that the spectrum curves had ordinates almost three times as great as would
normally have been expected for a magnitude 4.7 shock. In that paper it was
stated:
"The Port Hueneme aecelerogram is the first strong-motion recording of
a single-pulse shock. It is thus now known that such exceptional strong ground
motions can occur, at least for shocks up to 4.7 magnitude. This raises the
question whether it is possible that shocks of larger magnitude could be of
this type, for if a shock of magnitude 6 or greater could be of the single-pulse
type, it would require a revision of engineering thinking with respect to possible intensities of ground motions. On the basis of past experience, it would
appear that the probability of a strong single-pulse shock is relatively small
and, hence, engineers should continue to be guided by typical earthquake
ground motions. The possibility of strong single-pulse shocks must, however,
be kept in mind."
The Parkfield earthquake thus proves that a pulse-type shock can be generated
by a magnitude 5.6 event. This raises the question as to the possibility of a magnitude 6.5 or 7.5 shock generating an even larger pulse. In the Parkfield earthquake the slip apparently extended over a length of fault of some 20 miles and it
does not seem likely that a more powerful pulse would have been produced if the
slip had extended over a 40-mile length or over twice the depth. It woulci ~eein
that a greater stress drop on the fault would be required to generate a greater
pulse.
ENGINEERING SIGNIFICANCE
The fact that a displacement pulse of 10 inches amplitude can occur at a fault
is of special engineering significance. Important structures to be located close to a
fault along which an earthquake similar to the Parkfield shock may occur should
therefore, be designed to withstand a ]0-inch displacement pulse without serious
consequences.
Also of special engineering significance is the fact that earthquake ground motion of 40 to 50 per cent of g and spectral velocities of 50 to 80 inches per second
do not necessarily constitute a very destructive earthquake. The small damage
caused by this earthquake points up the fact that large ground accelerations and
large spectral values do not in themselves indicate severe damage to structures,
at least of the type in the Cholame-Parkfield area. Such motion, of course, might
be highly destructive of unreinforced masonry buildings. It should be noted that
except for the bridge there were no structures near Station 2 to receive the unmodified effect of the 10-inch pulse. The nearest structures were at Station 3 in
Cholame, and the spectrum here, as indicated by the seismoseope record, was
already reduced from that at Station 2.
The different shapes of spectrum eurves at locations not widely separated and on
similar local geology indicate that in this earthquake the shapes of the spectra
were determined largely by such factors as source mechanism and distance from
causative fault rather than by local geology.
1220
B U L L E T I N OF THE SEISMOLOGICAL SOCIETY OF AMERICA
T h e unexpected and very significant information revealed b y the Cholame array
emphasizes the desirability of installing additional such arrays at strategic locations across active faults. I t is clear t h a t had the Parkfield earthquake been recorded only at Station 12, ten miles from the fault, there would have been no
convincing evidence t h a t this shock was in a n y w a y significant, or t h a t there had
been large displacements and accelerations associated with the earthquake.
ACKNOWLEDGMENTS
The cooperation of the California State Department of Water Resources and the U. S.
Coast and Geodetic Survey in making available for study the accelerograms is acknowledged. The analysis of the accelerograms was supported in part by a grant from the
National Science Foundation Division of Engineering.
REFERENCES
Cloud, W. K. (1967). Intensity map and structural damage, Parkfield, California earthquake
of June 27, 1966. Bull. Seism. Soc. Am. 57, 1161-1178.
Cloud, W. K. and V. Perez (1967). Accelerograms-Parkfield earthquake, Bull. Seism. Soc.
Am. 57, 1179-1192.
Cloud, W. K. (1967). Seismocope results from three earthquakes in the Hollister, California.
area, Bull. Seism. Soc. Am. 57,
G. V. Berg and Housner, G. W. (1961). Integrated velocity and displacement of strong earthquake ground motion, Bull. Seism. Soc. Am. 51, 175-189.
Housner, G. W. and D. E. Hudson (1958). The Port Hueneme earthquake of 1V[arch 18, 1957,
Bull. Seism. Soc. Am. 48, 163-168.
Hudson, D. E. and W. K. Cloud (1967). Analysis of seismoscope data from Parkfield earthquake of June 27, 1966, Bull. Seism. Soc. Am. 57, 1143-t159.
D i v i s i o n OF ENGINEERIN(] AND APPLIED SCIENCE
CALIFORNIA INSTITUTE OF TECHNOLOGY
PASADENA~ CALIFORNIA
l~'Ianuseript received June 15, 1967.