of 6
TELESEISMIC
MECHANISM
OF
THE
MAY
2,
1983
COALINGA,
CALIFORNIA,
EARTHQUAKE
FROM
LONG-PERIOD
P-WAVES
by
Stephen
H.
Hartze11
1
Thomas
H.
Heaton
2
ABSTRACT
Teleseismic,
long-period
P-waveforms
are
modeled
to
obta
in
est i -
mates
of
the
source
parameters
for
the
May
2,
1983
Coalinga
earthquake.
The
best
fitting
focal
mechanism
is:
strike=
297
±5° ,
dip=
64
±1°,
rake=
70
±10°.
The
moment
is
estimated
to
be
3.8
±1 .5
X
10
25
dyne-e
m
with
a
slip
duration
of
5
±1
sec.
The
depth
is
est
im
ated
at
12
± 2
km.
1
.
Woodward-Clyde
Consultants~
P.O.
Box
93245,
Pasadena
,
CA
91109
2
u.s.
Geological
Survey,
California
Institute
of
Techno
l ogy,
Pasadena,
CA
91125
241
242
CALIFORNIA
DIVISION
OF
MINES
AND
GEOLUGY
ANALYSIS
AND
DISCUSSION
Long-period,
vertical
P-waves
from
WWSSN
and
Canadian
stations
in
the
distance
range
from
30°
to
goo
were
collected
for
the
Coalinga
earthquake.
The
P-waves
are
shown
in
Figure
1 (boldface
traces).
These
waveforms
are
modeled
using
the
generalized
ray
technique
(Helmberger
and
Harkrider,
1g78).
The
crustal
structure
used
in
the
modeling
is
given
in
Table
1.
In
choosing
this
velocity
structure,
the
results
of
refraction
surveys
by
the
U.S
.
Geological
Survey
(Walter
and
Mooney,
1g82)
were
used
as
guidelines.
Because
we
are
using
a layered
crustal
model,
the
ray
expansion
needed
in
the
gen
-
eralized
ray
method
is
complicated.
To
ensure
that
the
ray
expansion
contains
all
significant
rays,
the
results
were
checked
against
a
Haskell
matrix
code.
The
value
of
t*
is
assumed
to
be
1.0.
The
strike
and
dip
of
the
north-easterly-dipping
nodal
plane
(2g7o
and
64°,
respectively)
are
fairly
well
constrained
by
the
downward
first
motion
at
PHC
and
by
the
nodal
up
first
motions
at
ALE
and
WES.
By
fitting
these
key
stations,
the
first
motions
at
all
the
remaining
stations
are
satisfied
(for
a
thrust
mechanism).
Although
there
is
some
trade-off
between
the
assumed
velocity
in
the
source
region
and
the
fault
dip,
reasonable
perturbations
in
the
assumed
velocity
struc-
ture
results
in
variations
in
the
dip
angle
of
about
a degree.
Local
first-motion
data
suggest
a rake
of
near
goo
(Eaton,
1g83).
However,
the
teleseismic
P-waveforms,
particularly
those
at
MAT,
SKH,
and
WES,
appear
incompatible
with
a rake
of
goo
or
greater.
These
wave-forms
are
better
described
by
a rake
of
less
than
goo.
Figure
2 compares
waveforms
for
different
rakes.
We
have
chosen
a rake
of
70°
since
it
does
not
violate
many
of
the
local
first
motions
and
fits
the
tele-
seismic
P-wave
data
well.
However,
the
error
bars
are
relatively
large,
about
±10°.
The
corresponding
conjugate
nodal
plane
has
a
strike
of
156
±15°,
a dip
of
32±5°,
and
a rake
of
125±10°.
Since
we
approximate
the
source
with
a point
source,
we
cannot
choose
between
the
conjugate
planes
based
upon
this
analysis.
A comparison
of
the
synthetics
for
this
mechanism
with
the
data
is
shown
in
Figure
1.
The
best
fitting
average
moment
is
3.8
X
10
25
dyne-em.
All
the
stations
were
used
to
obtain
this
value
except
LPA,
which
has
an
anomalously
low
amplitude.
Amplitudes
of
the
synthetics
in
Figures
1
and
2
are
based
on
this
moment.
A very
similar
mechanism
has
been
obtained
by
Kanamori
(1g83,
this
volume)
from
surface
wave
data.
The
far-field
source-time
function
used
is
a trapezoid
(pictured
in
Figure
1)
with
a
rise
of
1
sec,
a top
of
1
sec,
and
a
fall
of
3 sec.
A 5
sec
total
source
duration
is
generally
consistent
with
the
duration
of
strong
shaking
recorded
by
the
Pleasant
Valley
Pumping
Plant
Station
g.4
km
from
the
epicenter
(Maley
et
al
.,
1g83).
Long-period
body
waves
are
not
particularly
sensitive
to
the
depth
of
the
source.
However,
we
can
set
some
upper
and
lower
bounds.
Most
of
the
P
wave-forms
in
Figure
1
have
a
positive
deflection
about
15
sec
after
the
first
arrival.
This
phase
is
made
up
of
crustal
multiples
of
Moho
reflections
involving
S
to
P
conversions
at
the
free
surface,
and
gives
some
control
on
the
depth.
SP66
1983
COALINGA
EARTH~UAKES,
MAY
1983
The
synthetics
in
Figures
1
and
2
are
calculated
for
a depth
of
12
km.
The
match
between
the
synthetics
and
the
data
degrades
for
depths
less
than
10
km
and
greater
than
14
km.
There
are
of
course
trade-offs
be-
tween
these
depths
and
the
velocity
structure.
Figure
3
shows
teleseismic
short-period
P-waves.
All
the
records
are
from
WWSSN
stations
except
ESK
which
is
a special
instrument
with
a response
similar
to
WWSSN.
The
records
are
lined
up
on
the
first
major
upswing,
which
can
be
identified
in
each
of
the
8 seismograms.
There
is
a
fairly
coherent
low-amplitude
arrival
about
two
seconds
ahead
of
this
first
major
peak.
There
appears
to
be
a
two
second
low-amplitude
initiation
phase, suggesting
less
energetic
faulting
at
the
beginning
of
the
rupture.
This
observation
is
consistent
with
initial
low-level
ground
motions
recorded
on
high-gain
digital
instruments
in
the
Parkfield
area
(William
Ellsworth,
Menlo
Park,
personal
communication).
A
similar
feature
in
the
teleseismic
short-
period
P-waves
for
the
1979
Imperial
Valley
earthquake
has
been
reported
by
Hartzell
and
Heaton
(1983).
CONCLUSIONS
The
mechanism
of
the
1983
Coalinga,
California
earthquake
determined
by
modeling
of
teleseismic
long-period
P-waves
is
found
to
be
strike
=
297°,
dip=
64°,
rake=
70°,
with
a
moment
of
3.8
X
10
25
dyne-em.
The
depth
is
estimated
at
12
km
with
a source-time-function
duration
of
5
sec.
The
emergent
character
of
the
teleseismic
short-period
P-waves
suggests
that
the
rupture
started
with
a 2 sec
interval
of
less
energetic
faulting.
ACKNOWLEDGMENTS
The
authors
are
grateful
to
the
operators
of
WWSSN
stations
who
supplied
records
for
this
study.
We
thank
Hiroo
Kanamori
for
use-
ful
discussions
and
for
reviewing
the
manuscript.
243
244
CALIFORNIA
DIVISION
OF
MINES
AND
GEOLOGY
REFERENCES
Eaton,
J.
P.
(1983).
Seismic
setting,
location
and
focal
mec
hanism
of
the
May
2,
1983,
Coalinga
earthquake,
in
Coalinga
earthquake
sequence
commencing
May
2,
1983,
U.S.
Geological
Survey
Open
File
Report
83-511.
Hartzell,
S.
H.
and
T.
H.
Heaton
(1983).
motion
and
teleseismic
waveform
data
history
of
the
1979
Imperial
Valley,
Bull
Seism.
Soc.
Am.,
December.
Inversion
of
strong-ground
for
the
fault
rupture
California
earthquake,
Helmberger,
D.
V.
and
D.
Harkrider
(1978).
Modeling
earthquakes
with
generalized
ray
theory,
in
Modern
Problems
in
Elastic
Wave
Propagation,
J . Miklowitz
and
J.
D.
Achenback,
Editors,
John
Wiley
and
Sons,
New
York.
Kanamori,
H.
(1983).
Mechanism
of
the
1983
Coalinga
earthquake
determined
from
long-period
surface
waves,
CDMG
Spe
c
ial
Report,
The
1983
Coalinga,
California,
earthquake.
Maley,
R.,
G.
Brady,
E.
Etheredge,
D.
Johnson,
P.
Mark,
J.
Switzer
(1983).
Analog
strong
motion
data
and
processed
main
event
records,
in
Coalinga
earthquake
sequence
commencing
May
2,
1983,
U.S.
Geological
Survey
Open
File
Report
83-511.
Walter,
A.
W.
and
W.
D.
Mooney
(1982).
Crustal
structure
of
the
Diablo
and
Gabilan
ranges,
central
California:
a
reinterpre-
tation
of
existing
data,
Bull.
Seism.
Soc.
Am.,
I£,
1567-1590.
Table
1
Velocity
Structure
~
(km/sec)
8
(km/sec)
p
{gm/cm
Th
(km)
3.0
1.70
2.1
3.0
5.5
3.17
2.5
3.0
6.2
3.58
2.7
20.0
8. 0
4.62
2.9
SP66
1983
COALINGA
EARTHQUAKES,
MAY
1983
COALINGA
Vertical
P
waveforms
M
0
=
3,8xi0
25
dyne-
em
Strike
=
29?D
Dip
=
64°
Rake
=
70°
0
20
40
sec
Peak
amp
lit
udes
given
in
microns
A~.9
f--!-·'
BOG
7.3
~
~
LPS
7.
1
~
~8
~
11111
3
I
MAT
7.3
~
~2
v~.s
~-7
PEL
55
~-
~-'
Figure
1.
Comparison
of
observed,
vertical,
long-period,
tele-
seismic
P-waveforms
(boldface
traces}
with
generalized
ray
synthetics
for
the
May
2,
1983
Coalinga
earthquake.
The
focal
mechanism
is
an
equal
area
projection
.
245
246
CALIF
O
RNIA
DIVISION
OF
MINES
AND
GEOLOGY
w
=--:.
{\-..
A
::_
5
Obs
.
-v
vv-
700
~
9
0°~
110°~
9
MAT
7 . 3
~
~
~
~
0
20
40
sec
SHK
7 . 9
~
~9
~
6
~
6
Figure
2.
Comparison
of
P-
waveforms
for
rakes
of
70°,
90°,
and
110°.
Am~litudes
are
given
in
microns
for
a
moment
of
3.8
X
10
2
dyne
-
em.
A
KU
Short
Period
P
Waves
I
I
I
I
0.13
j
ILL
ALE
-rwv~w
0 .19
CAR
ESK
I
I
I
I
I
I
0 .14
0.17
(Spec
i
al)
NU
R
~
0.
29
I
0
10
Sec
Figure
3.
Short-peri
od
te
les
eis
mic
P-waves
.
These
data
s uggest
a
two
seco
nd
initiation
phase
of
low-amplitude
faulting
.
Amplitudes
are
given
in
microns.
SP66