of 10
TILTS
ASSOCIATED
WITH
THE
PT.
MUGU
EARTHQUAKE
R.
W.
Alewine,
III
and
Thomas
H.
Heaton
Seismological
Laboratory
California
Institute
of
Technology
Pasadena,
California
91109
Abstract.--Anomalously
large
far
field
tilts
were
observed
during
the
Pt.
Mugu
earthquake
of
February
21,
1973.
The
magnitude
and
pattern
of
tilting
is
not
explained
by
a
simple
elastic
dislocation
model
using
reasonable
source
parameters
for
this
event.
Dynamical
triggering
of
small
displacements
on
a
fault
well
exterior
to
the
epicentral
region
is
offered
as
an
alternative
explanation
to
the
observed
large
tilts.
Observations.--A
total
of
nine
tiltmeters
were
operating
at
seven
southern
California
stations
during
the
occurrence
of
a
local
magnitude
6.0
earthquake
near
Pt.
Mugu
on
February
21,
1973.
Figure
1
shows
the
location
of
the
epicenter
of
this
event
and
the
station
distribution
in
relation
to
the
active
fault
systems
in
the
area.
The
tiltmeter
at
station
ISA
was
designed
by
Hugo
Benioff
and
consists
of
two
mercury
pots
separated
by
a
thirty
meter
tube.
Tilts
are
recorded
by
sensitive
differences
in
the
relative
levels
of
mercury
in
each
container.
A
smaller
field
model
of
this
instrument
was
in
operation
at
stations
SSR
and
MHSa.
The
instruments
at
stations
PAS,
HAG,
and
MHSb
are
long
period
horizontal
garden-gate
type
seismometers
with
a
displacement
sensitive
transducer.
This
type
of
instrument
has
been
described
by
Gilman
(1960).
Further
discussions
of
the
tiltmeters
operated
by
the
Caltech
Seismological
Laboratory
is
given
by
Gile
(1973).
As
a
check
on
the
relative
response
and
drift
characteristics
of
the
instruments,
a
Gilman
tiltmeter
(garden-gate)
was
placed
in
operation
parallel
to
a
Gile
tiltmeter
(mercury
pot)
at
station
MRS.
Two
of
the
stations,
ISA
and
PAS
are
operated
on
a
permanent
basis
in
granite
tunnels,
while
stations
HAG,
SSR,
and
MRS
are
temporary
stations
located
on
fanglomerate
along
the
Raymond
fault
near
Pasadena.
The
instruments
at
the
temporary
stations
were
calibrated
directly
by
means
of
a
leveling
screw.
In
addition
to
these
instruments,
a
mercury
pot
type
tiltmeter
designed
by
Alfred
Focke
of
Harvey
Mudd
College
(HMC)
and
a
bubble
tiltmeter
built
by
Autonetics
(NAR)
were
in
operation
at
the
time
of
the
earthquake.
The
station
HMC
is
located
on
a
fanglomerate
type
material
several
hundred
meters
from
the
Cucamonga
fault.
Contribution
No.
2372,
California
Institute
of
Technology,
Division
of
Geological
and
Planetary
Sciences,
Pasadena,
California
91109.
94
tN
0
10
20
30
40
50
km
Santa
Barbara
Channel
EPICENTER
PT.
MUGU
EARTHQUAKE
SSR
Cucamonga
E
p~~:~2.--
-:.::..;~-==:::::.-=~
----
HAR
MHS
HMC
Los
Angeles
Figure
1.--Tilt
station
distribution
in
relation
to
epicentral
location
and
tectonic
framework.
Pertinent
data
for
the
eight
stations
and
the
magnitude
and
direction
of
observed
tilt
are
given
in
Table
1.
The
sign
of
the
tilt
is
determined
by
the
following
convention.
A
positive
tilt
indicates
that
either
north
or
east
is
upward
relative
to
south
or
west
respectively.
In
this
table,
stations
which
have
instruments
aligned
off
north-south
or
east-
west
have
the
observations
rotated
into
this
coordinate
frame.
Presented
in
Figure
2
are
examples
of
the
records
of
observed
tilting
at
the
temporary
stations
during
this
event.
In
each
tilt
recording,
there
is
no
indication
of
any
precursory
tilt
before
the
arrival
of
the
P-wave
from
the
hypocenter
of
the
main
shock.
The
static
tilt
appears
to
reach
its
final
value
within
a
period
of
about
three
minutes
of
the
onset
of
the
dynamical
motion.
The
slow
recording
speed
prevents
further
resolution
of
the
time
interval
necessrary
to
reach
the
static
value.
The
static
tilt
did
not
recover
within
a
period
of
at
least
three
hours,
but
we
were
unable
to
95
Station
PAS
ISA
HAG
SSR
MHSa
MHSb
HMC
NAR
Dist.
from
Epicenter
(km)
77
201
81
84
91
91
132
88
Table
1
Geologic
Setting
Granite
tunnel
Granite
tunnel
Fanglomerate
Fanglomerate
Fanglomerate
Fanglomerate
Fanglomerate
Alluvium
Comp.
of
Tilt
E-W
N-S,E-W
N-S
N-S
E-W
E-W
N-S
N-S,E-W
Offset
(ll
radians)
0.056
<0.01
6.0
--10.0
-5.0
-6.0
3.0
<0.5
Figure
2.--Sample
of
observed
tilts
at
the
temporary
stations
located
along
the
Raymond
Fault.
Arrows
in
figures
a
and
b
indicate
the
occurrence
of
a
local
"after
event."
96
determine
or
not
if
the
tilts
were
truly
permanent
due
to
the
drift
in
the
instrument.
In
this
figure,
a
small
long
period
event
can
be
seen
on
both
MHSa
and
b
about
two
hours
after
the
main
event.
There
is
no
seismic
activity
within
an
hour
of
that
time
which
was
recorded
by
the
Southern
California
seismometer
array.
Although
this
small
event
is
not
seen
on
HAG
and
SSR,
it
might
be
interpreted
as
a
local
"after-event"
to
the
larger
tilt
event
which
occurred
during
the
Pt.
Mugu
earthquake.
The
magnitude
and
direction
of
the
tilt
steps
recorded
on
the
instruments
running
in
parallel
at
MHS
were
in
excellent
agreement,
significantly
increasing
our
confidence
in
the
observed
data.
It
should
be
noted
that
all
four
stations
at
which
large
tilt
steps
were
measured,
are
located
on
fanglomerate.
However,
no
tilt
step
was
seen
at
NAR
which
is
also
located
on
thick
alluvium.
One
would
expect
that
simple
settling
would
occur
in
such
a
way
that
the
local
topographic
gradient
of
the
area
would
be
lessened.
However,
the
tilt
offsets
consistently
had
the
effect
of
steepening
the
local
topographic
gradient.
Dislocation
modelling
of
observed
tilts.--From
the
focal
mechanism
studies
of
Whitcomb
(1973)
using
only
first
motion
data
of
P-waves,
from
stations
near
the
epicentral
area,
three
different
source
mechanisms
are
possible
for
this
event.
The
tilt
fields
due
to
each
of
these
mechanisms
were
computed
using
the
integral
expressions
for
strains
due
to
a
static
dislocation
in
a
uniform
half-space
given
by
Canitez
and
Toksoz
(1972).
These
expressions
were
numerically
integrated
over
the
appropriately
oriented
fault
plane
for
each
of
the
two
examples
studied.
Since
the
areal
extent
of
the
fault
dislocation
surface
could
only
be
estimated
by
the
size
of
the
aftershock
zone,
we
assumed
a
somewhat
exaggerated
fault
dimension
in
order
to
get
a
lower
limit
of
the
calculated
dislocation.
In
the
calcula-
tions,
we
assumed
that
the
dislocation
was
distributed
uniformly
over
a
surface
extending
from
2
km
below
the
surface
down
to
the
hypocentral
depth
of
20
km,
and
had
a
length
of
8
km.
The
results
of
these
calculations
are
given
in
Table
2
below.
The
calculated
tilts
are
given
in
units
of
~-radians
per
meter
of
dislocation
on
the
fault
surface.
The
tilts
have
been
rotated
to
the
orientation
of
the
instruments.
Strike
N
69
E
N
63
E
DIP
41°NW
39°NE
PAS
.02
.006
Table
2
ISA
-lo-s
-4
-10
HAG
.02
.007
97
SSR
MHS
.02
.009
• 007
.
0(16
HMC
NAR
.006
-.005
.003
.005
It
is
seen
from
this
table
that
even
in
the
most
favorable
of
cases,
Case
1,
for
the
most
reliable
tilt
observation,
that
at
PAS,
a
minimum
of
300
em
of
dislocation
on
the
exaggerated
fault
surface
is
required
to
explain
this
observation.
This
value
for
the
fault
dislocation
is
at
least
a
factor
of
10
greater
than
that
required
by
Kanamori
(1972)
to
satisfy
the
observed
dynamic
data
at
PAS
using
Haskell's
(1969)
dislocation
model.
For
the
other
temporary
stations,
more
than
two
orders
of
magnitude
more
of
dislocation
is
necessary
to
explain
their
respective
readings.
The
estimate
of
the
fault
dislocation
goes
up
even
more
if
a
more
reasonable
source
dimension
for
this
size
earthquake
is
assumed.
Hanks
(1973)
has
determined
the
moment
from
the
long
period
spectral
level
to
about about
1.2
x
10
24
dyne-em,
or
about
two
orders
of
magnitude
smaller
than
that
needed
to
fit
the
recordings
at
PAS,
SSR,
MHS,
and
HMC.
Since
Jungels
and
Anderson
(1971)
were
able
to
obtain
satisfacotry
results
in
fitting
far
field
strain
measurements
with
an
even
more
crude
static
fault
model
in
the
case
of
the
San
Fernando
earthquake,
1971,
we
assume
that
the
simple
dislocation
model
used
here
can
adequately
explain
the
strains
to
at
least
an
order
of
magnitude.
If
this
is
true,
then
the
tilts
observed
can
not
be
due
to
the
direct
far-field
static
strain
field
caused
by
the
main
shock
in
the
Pt.
Mugu
sequence.
Also,
although
stations
PAS
and
MHS
are
on
approximately
the
same
azimuth
from
the
epicenter,
they
show
opposite
directions
of
tilts.
As
can
be
seen
from
Figure
3a,
the
far-field
E-W
tilt
field
for
Case
1,
the
most
likely
solution,
shows
no
node
within
30°
of
this
azimuth.
Figure
3b
shows
the
far
field
N-S
tilt
field
for
this
case.
Both
the
temporary
stations
and
PAS
are
near
a
node
in
this
field
demonstrating
the
poor
solution
to
the
problem
this
simple
model
provides
of
the
direct
strain
field.
One
possible
explanation
for
the
large
strains
observed
is
afforded
if
a
small
dislocation
is
"triggered"
at
some
distance
away
from
the
source
by
the
passage
of
elastic
waves
through
some
prestressed
area.
A
documented
example
of
this
occurrence
is
given
by
Allen
et
al.
(1972)
for
the
case
of
the
1968
Borrego
Mountain
earthquake.
In
this
example,
displacements
of
one
to
two
and
a
half
centimeters
over
tens
of
kilometers
of
fault
were
triggered
at
a
distance
some
fifty
to
seventy-five
kilometers
away
from
the
main
event.
The
stations
at
which
anomalous
tilts
were
observed
for
the
Pt.
Mugu
earthquake
all
lie
along
the
fault
zone
which
forms
the
southern
boundary
of
the
Transverse
Range
Province.
This
zone
contains
the
Malibu
Coast
Fault,
the
Santa
Monica
Fault,
the
Raymond
Fault,
the
Sierra
Madre
Fault,
and
the
Cucamonga
Fault.
In
this
fault
zone,
the
offset
has
been
a
combination
of
left-lateral
strike-slip
and
reverse
thrusting.
Since
the
stations
where
anomalously
large
tilts
were
recorded
were
located
near
the
Raymond
fault,
we
chose
to
try
to
explain
the
large
strains
by
a
triggered
dislocation
along
this
fault.
This
triggering
could
be
possible
only
if
the
area
around
the
fault
surface
were
in
a
relatively
high
prestressed
state.
98