JOURNALOl GEOPHYSICALR.s.

Aacn

VOL. 72, No. 4

FEBaUAaY 15, 1967

Scaling Law of SeismicSpectrum

K,iii

Axi

Department o Geology and Geophysics

MassachusettsInstitute oi Technology, Cambridge

The dependenceof the amplitude spectrumof seismicwaves on sourcesize is investigated

on the basis of two dislocationmodels of an earthquake source.One of the models (by N.
Haskell) is called the o8 model, and the other, called the 2 model, is constructedby fitting
an exponentially decaying function to the autocorrelation function of the dislocation velocity.
The number of source parametersis reduced to one by the assumptionof similarity. We
found that the most convenientparameter for our purposeis the magnitude M,, defined for
surfacewaves with period of 20 sec. Spectral density curves are determined for given M,.
Comparisonof the theoretical curves with observationsis made in two different ways. The
observedratios of the spectra of seismicwaves with the same propagation path but from
earthquakesof different sizesare comparedwith the correspondingtheoretical ratios, thereby
eliminating the effect of propagation on the spectrum. The other method is to check the
theory with the empirical relation between different magnitude scales defined for different waves at different periods.The 2 model gives a satisfactoryagreementwith such observations on the assumptionof similarity, but the a model doesnot. We find, however,some
indicationsof departure from similarity. The efficiencyof seismicradiation seemsto increase
with decreasingmagnitudeif the Gutenberg-Richtermagnitude-energyrelation is valid. The
assumptionof similarity implies a constantstressdrop independentof sourcesize. A preliminary study of Love waves from the Parkfield earthquake of June 28, 1966, shows that the
stress drop at the source of this earthquake is lower than the normal value (around 100 bars)
by about 2 orders of magnitude.
INTRODUCTION

longer-periodwaves are generated.In the early

days of seismologyin Japan, much attention
years to find seismicsourceparameterssuchas was given to the presenceof large long-period
fault length, rupture velocity,and stressdrop motion in P wavesfrom large local earthquakes
at the earthquakesourcefrom the spectrumof [cf. Matuzawa, 1964]. Analysesof seismicwaves
seismicwaves.Except for the geometricparam- by Jones[1938], Honda and Ito [1939], Gtteneters obtainedfrom fault plane studies,how- berg and Richter [1942], Byerly [1947], Kanai
ever,the magnitudeis the only physicalparam- et al. [1953], Asada [1953], Aki [1956], Kasaeter that specifiesmost earthquakes.A gap hara [1957], Matumoto [1960], and others
exists between the two approachescurrently have shownthat the period of the spectralpeak
used,onebasedon the useof spectrumand the for P waves,S waves, surfacewaves,and even
otheron amplitude.The purposeof the present for codawaves increaseswith earthquake magElaborate

studies have been made in recent

paper is to fill this gap by finding a first

nitude.

The most convincingevidencefor the greater

efficiencyof generating long-period waves by
basisof somedislocation
modelsof earthquake larger earthquakesis probably given by Bercksources.For this purpose we must reduce to hemer [1962]. He comparedseismogramsobone the number of parametersspecifyinga tained at a station from two earthquakesof the
approximation to the relation between seismic
spectrumand magnitudeof earthquakeson the

dislocation.
We shall make this reductionby same epicenter but of different size. We shall
assumingthat large and small earthquakes reproducehis result later.
satisfy a similarity condition.
The relation betweenseismicspectrumand

The magnitudeof an earthquakeis definedas

a logarithm of amplitude of a certain kind of
earthquakemagnitudeis not a new problem. seismicwave recordedby a certain type of
It has been well known that the greater the band-limited seismograph.If there is such a
size of an earthquake, the more efficiently size effect on seismic spectra as mentioned
1217

1218

KEIITI

above,the unit of magnitudeobtainedfrom one

kind of wave recordedby one type of seismograph may not correspondto that obtained

AKI
1

u0 - 4rbr
cos
20sin

from another kind of wave recorded on another

,t -- r- /cs

instrument.In fact, Gutenbergand Richter

[1956a] discovered
a discrepancy
betweenthe where a and b are the velocities of P and S
magnitudescalebaseduponshort-period
body waves,respectively.The aboveexpressions
have
wavesand that baseduponlong-periodsurface the foliowingcommonform:
waves.

It will be shownthat the theoretical

scaling V= P(r, O,,i:,,a, b)

law of the seismicspectrumderived from a

dislocationmodel of the earthquakesource
satisfactorilyexplainsthe above-mentioned
observations.

15 ,t

(3)

where c is the appropriatewave velocity.In

terms of the Fourier transform,the aboveform

THEORETICAL
MODELSOF THE EARTHQUAKE

can be written as

SOURCE

U(o)= P(r, O,'i:',a, b)A(co)

FollowingHaskell [1966], we definea dislocation functionD(, t) which is the displacement discontinuityacrossa fault plane at a
point and time t. The fault plane extends
along the / axis, and D(, t) is consideredas
the average dislocationover the width w of the

(4)

where

= u(O,at
A(o)
=

dt

fault. Taking the starting point of the fault at

the originof the (x, y, z) coordinates,
and the

,t -r--csO
axis along the x axis, we assumethat the
fault endsat = L and the surrounding
medium is infinite, isotropic,and homogeneous.If the medim is dissipative,the equationcorIntroducingpolar coordinates
(r, O, ,) by the respondhgto (4) will be
relation

U(oo)
= Pit, O,, a, b,oo,Q(o)].
A(o) (6)
x -'r

cosO

y = r sin 0 cosv

(1)

where Q(o) is the dissipationcoefficient.The

above expressionshowsthat .wecan isolatethe
propagation
term P((o) whichdoesnot, except

for the directionof fault propagation,

include
the fault motionparameters.Mathematically,
the displacementcomponentsof P and $ waves sucha simpleisolationis not permittedfor an
at long distances,corresponding
to a sourceof arbitrary heterogeneous
medium.Practically,
longitudinalshear fault [Haskell, 1964] for however,this separationof propagationfactor
example,can be written as
from sourcefactormay be permitted,at least
as a goodfirst approximation.
In this paperwe
z = rsin 0sinv

shall be concernedonly with the sourcefactor

A((o), which can be calculatedfrom the disloca-

U,.-- 4rbr sin20sin

0)
'wfo ,t--r- cos
a

tionD (, t) according
to (5).
For comparisonwith observationswe shall

use seismicwavesobservedat a given station

fromdistantearthquakes
of thesameepicenter,
thesamefocaldepth,andthesamefaultplane

cos 0 cosqo

solution,but of differentmagnitude.The ratio

Wfo

r- cos
b

of the Fourier transforms of two such seismo-

grams may be directly comparedwith the

SCALING

LAW

OF SEISMIC

theoretical ratio for the source factor A(o),

becausethe propagation factor P((o) may be
canceledin the observed ratio. This ingenious
method comes from Berckhemer [1962]. His
theoreticalmodel,however,seemsunrealistic,because,from the point of dislocationtheory, his
model impliesthat the amount of dislocationis
constant,independentof the size of the fault.
Following the generalline of approachtaken
by Haskell [1966],we introduce the autocorrela-

SPECTRUM

1219

(, ) = l.ff_
b(,0 ff B(,
e(+')-+) dk &od dt

'
I ff:B(k,o)B(--k--o

4-:

lB(k,o.,)l:

4w2

tionfunction(/, ) of (, t)'

dk &o

dk &o

Comparingthis formula with (9), we get the

(7)

well-known relation

(k, o)= lB(k,o)[

'

Putting the Fourier transform of k(, ') as

(k, o),weget

=ff

(14)

Finally, we get from (9), (13), and (14) the

relationbetweenthe amplitudespectraldensity

)e
-'+'"
dd. (8)IA(o)]and the Fouriertransformof the autocorrelation

(,, ) =

(k,w)e'
-'" dk (9)

On the other hand, A(w) can be retten by

changingthe order of integration and putting

function'

IA(w)["= w[(w cosO)/c,w]

(15)

Thus,the sourcefactor]A(co)]of the amplitude

spectral density is expressedin terms of the

t' = t --(r -- cosO)/cin (5) as follows'

autocorrelation
of dislocation
velocity/)(, t).
We followed Haskell [1966] in deriving the

A()
=we
-'"fffb(e,

expressions
above. Haskell, however,calculated
the energyspectraldensityfrom the autocorrelation function$(7, ) of dislocationacceleration

.e-.,+.o/ dt d

(10) /(, o.

In the above eression the integration fits

e extended
to ity

by putting(, t) = 0

for ( 0 and L ( . Puttg the Fourier trans-

(16)

formof (, t) asB(k, w),weobtain

The Fouriertransform
$(k, w) of ts function

s(,)=ff b(e,
t)e
-'"'*'
dtde
12ff
-

isrelated
to (k, w)simply
by
()

$(, ) = f(, )

(7)

Thus, we obtn
2

Then we have from (10) and (11)

' ,w

()1 = $[( o0)/,]

(12)

and

I.a()[: = w'

c , o

(13)

On the other hand, we get from (7) and (11)

0s)

As shownabove,the amplitudespectraldensity of seisc wavescan be expressedin te

of theautocogelation
function
of (, t) orthat
of (, t). The autocogelationfunction of
(, t) canbe detersnedif the absoluvalue
of the Fouriertramformof (, t) is ven.
There e an ite
number of space-time
fctio
that ve a commonspectrMdeity

1220

KEIITI

AKI

but have different phases.By specifyingan

autocorrelation,therefore, we are considering
an infinite group of space-timefunctions.The
model basedupon the autocorrelationfunction

//

is different from the deterministic one in this

respectand may be called 'statistical,'as was

doneby Haskell [1966].
Sincethe earthquakeis essentiallya transient
phenomenon,however,the autocorrelationfunc-

-T

//////
x%
0

tion introduced here cannot be treated in the

samemanner as the one for the stationarytime

series.The following figureswill schematically
i]lustrate what form may be expectedfor the
autocorrelation

function

for

the

I!

dislocation

processat an earthquake source. Let the dis-

locationstart at - 0 and propagatealongthe

axis with a constantvelocityv; then the dislocationat will be zero for t /v and will
take a constantvalue D0() for t T q- /v.
Figure I showsa schematicpictureof D(, t) at

a given. Thecorresponding/)(,
t) and(, t)

-T

"

"/

_ T
,

z:

II
II

Ii

Ii

Ii

Fig. 2. Schematic diagram of autocorrelation

are alsoshownin Figure 1. Their autocorrelation functions of dislocation velocity and dislocation
functionsare shownschematicallyin Figure 2. accelerationat a given point on a fault.
The dashedlinesin thesefiguresare for the case
in which the dislocation takes the form of a
In ourfirst model,weassumethat the temporal
ramp function in time. We now construct two autocorrelationfunction of dislocationvelocity
earthquake source models by fitting'simple decreases
exponentiallywith the lag r, that is
formulas to the two autocorrelation functions.

f_b(,t)b(,
tq-r)at- oe
-kr'l(19)
Our secondmodel is the one proposedby

O(S.t)

r'

Haskell. I-Ie assumes that the autocorrelation

function of dislocation acceleration takes the

followingform'
/,,

b (%.t)

(20)

We shall assumean identicalspatial correlation fction

>t

for both models. The correlation

betweenthe dislocation
velocityat $ and t and

that at $ + v andt' - t + V/v, that is

II

ff.D(e,
t)D(e
+

',,

de

in, cate the degreeof pendency of fat

propagation.The perstency1 decreaseth
ii

Fig. 1. Schematicdiagram of dislocationand its

time derivativesat a given point on a fault.

thedistance
v between
thetwopoints.FoHong
askell, we shahadopt the functionalform of
e- , for ts expression
,andso for the coe-

sportingfunction
of (, 0.

SCALING

LAW OF SEISMIC

The above temporal and spatial autocorrela- we have

tion functions are expressedin a single form, if
we write

SPECTRUM

1221

%/4krkro
k,kL

= wDoL

(29)

Inserting this into (25), we get

(21) IA)I

= oe-'"'-'-"/"
for the first model,and

wDoL

__

1+ co_s
c 0 2 o2 {lq-(w/k)2}
(30)
for our first model.

e-'-""
for the second model. Their Fourier

(22)
transfos

are

4k,kL,o

Since the above function

decreasesproportionallyto oJ-'-for large oz,we

shall call this the 'o>squaremodel.'
On the other hand, the sourcefactor of amplitude spectral density for our secondmodel will
decrease proportionally to o- for large oJ.

q0is equalto LD dkrk ,/8, according

to Haskell.

(,) = / + ( --/)}(+

Inserting this into (24) and (28), we obtain:

(23)

(,) = /:+ (8k'kL'qa'2

--/)}(+
(a)
Using (15) and (23), we may obtain the
sourcefactor of amplitude spectral density for
our first model as follows'

wDoL

)('-')}
202/2
{i-q-(
co-so
{1-'()
:}
(31)

We shall call this the 'o>cube model.'

ASSUMPTION OF SIMILARITY

w/4k,kL
o

The straightforwardway of testing the earthquake sourcemodelsproposedabove is to compare the predicted spectrum directly with the
observed one. For this purpose, however, we
must know about such effects of the propagation medium as dissipationand complexinterferenees on the seismic spectrum for a wide
frequencyrange. Although such knowledgehas
been accumulating,especially for long-period
waves [ef. Press, 1964], it does not yet satisfactorily cover the frequency range required
for the presentstudy.
As mentioned in the preceding section, we
will removethis difficulty by comparingseismic
waves having a common propagational path
but coming from earthquakesof different sizes.
Further, in order to specify an earthquakeby
a singlesourceparameter,'magnitude,'we must
reduce to one the number of parameters appearingin (30) and (31) by assumingthat they

[k2
+/.COS
0 )2(,2
]1/2+ (.1,)2)
1/2
To determinethe value of o, we put o = 0 in
(10). Then

(26)

=w
Comparingthe aboveequationwith (25), we get

%/4krkL
o= w

(27)

If we define an averagedislocationby

Oo= Z

(28)

are related to each other in some manner.

1222

KEIITI

The simplest of such assumptionsmay be

that large and small earthquakes are similar
phenomena.If any two earthquakesare geometrically similar, the fault width w is proportional to the length L. If they are physically
similar, all the nondimensional
productsformed
by the sourceparameterswill be the same.The
averagedislocationDo will be proportionalto L
and, consequently,to w. This implies that if an
earthquakeis a Starr fracture, the pre-existing
stress or strength is constant and independent
of sourcesize [Tsuboi, 1956]. Since the wave
velocity is practically independent of source
and may be consideredconstantfor our present
purpose,all the quantitieshaving the dimension
of velocity must also be constant and independent of source size. Thus, the similarity
assumptionsimply that the rupture velocity v
is a constantand that all the quantitieshaving
the dimensionof time, suchas k- and (vkL)-,
are proportionalto L.
For simplicity, we shall further assumethat

AKI

y(t)-- w

(32)

'" 42

whereoois given by the equation

t - --(d$/dco)
....

(33)

If this approximationis valid, the trace amplitude of waves with frequency read directly
on the recordwill be proportionalto the spec-

tral density]Y()I.

The quantityd/&o" in

(32) is the sum of a propagation term and a

sourceterm. Sincethe propagationterm is proportional to the travel distance, the source
term may be neglectedat long distances.Thus,
we may assume that the trace amplitude of
surface waves with period of 20 sec is equal
to the amplitude spectral density of waveswith
cos 0 -- 0 and that vkL -- k. A value of k
that period, except for a factor that is indegreaterthan vk may be a more realisticchoice, pendent of the sourcesize. The validity of this
becausek - is related to the time required for assumptionis confirmedby comparingthe ratio
formation of fracture across the fault width, of traceamplitudes
of Lovewaveswith a cerwhereas (vk) - is related to the time required tain period from two aftershocksof the Kern
for propagationof fracture along the length of County earthquakewith the ratio of amplitude
the fault. We shall examine later the case in
spectral densities at that period obtained by
which 10 vk -- kr. Essentiallythe sameresult the Fourier analysismethod. Both ratios agree
as when vk -- kr will be obtained,exceptfor well.
the value of k correspondingto a sourcesize.
Thus, the dependenceof amplitude spectral
SCALING Lw OF SEISMIC SPECTRUM

Under the assumptionsdescribedin the precedingsection,we can expressthe sourcefactor

of amplitude spectraldensityas a function of
L, % and several nondimensionalconstants.
Taking L as a parameter,we shall obtain a
group of curves of spectral density, each of
whichcorresponds
to an earthquakeof a certain
size. In order to find which curve corresponds
to a given earthquakesize,we must have a scale
to measure size. The most convenient scale for

density,IA()I, on the magnitudeM will be

suchthat log IA(o)]at the periodof 20 secis
equal to M plus a constant.In other words,
two spectrum curves corresponding to two
earthquakesizesdiffering by M -- 1.0 will be
separated by 1.0 along the ordinate at the

periodof 20 sec,if the curveIA()I is drawn

on a logarithmic scale. Figures 3 and 4 shows
such groups of curves for the o-squareand
-cube models,respectively.
The

curves shown in each of these charts

have an identical shape. The frequency that

our purposeis the surfacewavemagnitudescale, characterizesthe shape of the curve, such as
definedby Gutenbergand Richter [1936]. This k, is proportionalto L -, and the spectraldenmagnitude,designatedas Ms, is proportionalto sity at -- k is proportionalto L ', as can be
the logarithm of amplitude of teleseismicsur- found from (30) and (31) under the assumpface waves with period of about 20 sec. Since tion of similarity. Therefore, the points corat this period the waves are usually well dis- respondingto the characteristicfrequencylie
persed,we may expressthe wave train y(t) by on a straight line with gradient 3, as shownby
the stationaryphaseapproximation,as follows: dashedlines in Figures 3 and 4. As mentioned

SCALING LAW OF SEISMIC SPECTRUM

PERIOD
0.1

0.2

0.5

tO-SQUARE

IN
5

I0

SEC
20

50

I00 200

nitudes.The magnitudeof earthquakesstudied

by him coversthe range 4.5 to 8. After several
trials, we choosethe absolutevalue of magnitude that gives the best agreement between
theory and observation.The valuesassignedto
the curves in Figures 3 and 4 are determined
in this manner, and the correspondingtheoretical spectral ratios are shown in Figure 5,
together with the observedratios given by

500 I000

MODEL

1223

'

Berckhemer.

LOVE WAVES FROM,Two CALIFORNIA SHOCKS

The applicability of the theoretical curves

of spectral densitiesobtained in the preceding
sectionis tested by the use of recordsof Love

waves

f-,/

from

two

aftershocks

of

the

Kern

7.0 County, California, earthquake of

8.5

1952. The
epicentersof these two earthquakesare within
severalmiles of each other, accordingto Richter
[1955], and they show identical first motion
patterns, according to Bdth and Richter
PERIOD

0.5

03-CUBE

IN

I0 20

SEC
50 I00 200 500 I000

5(300

MODEL

M,defined

_ 8.o

FREQUENCY IN C/S
,

Fig. 3. I)eperderceof amplitude spectra] density of earthquake magnitude M, for the -square
model.

'"'

/f --- 7.5

before,the spacingof curvesfor differentearthquake magnitudesis determinedby the deft-

6.5

nition of M,. The definitionalone,however,

cannotgivethe absolute
valueof magnitude

,
'Y' ,.o

corresponding
to eachcurve.
If

we know the absolute value

for one of

the curves,the values for the rest are determinedfrom the definitionof M. First we

4.5

adopta trial valueof magnitude

for one of
the curvesand assignmagnitudevaluesto
other curves accordingto the definition. Then
we can find the ratio of spectral densitiesfor
two different magnitudesas a function of frequency or period. This ratio is comparedwith
the observedone given by Berckhemer [1962].

2,0

1.0 0.5

02

0.1 0.05

FREQUENCY

02

0.010005

IN

OJ:)02

0.0006

C/S

Fig. 4. Dependence.of amplitude spectral den-

letsdatainclude
sixsetsof twoearthquakes
sityonearthquake
magnitude
M, forthee-cube

with the same epicenter but of different mag-

model.

1224

KEIITI

AKI

AI /A2!

8/6,5

IOO

..

AI/A21

80 -

8/7,5

60

40
/

20'

i0

20

30

40

'-m

50 s

IO

20

30

40

50

60

7,5

AI/A2

' -- T
70 S

AI/A21

7,4/6,5

50

6,5

20

20

I0

IO

20

30

I0

40

Ai/A2

20

30 S

AI/A21

300

200
2

6,2

IOO

I
5

I
IO

I
15

=T
20 S

I
5

I
IO

I
15

5,7

I
20 S

Fig. 5. Comparisonof theoretical and observedspectral ratio, plotted against period, for
pairs of earthquakeshaving nearly the same epicenterbut different size. Observedvalues are
reproducedfrom Berclchemer[1962]. The numbers shown for each pair are the earthquake
magnitudefor the pair. Solid line denotes-square model; dashedline denotes-cube model.

[1958]. The Richter magnitude(M,; localscale are obtainedby Berckhemer'smethod,in which

for southernCalifornia) of one of them is 6.1, the ratio is obtainedbetweenthe corresponding
and that of the other is 5.8. The difference of
peaks of waves by directly reading amplitudes
0.3 correspondsto the maximum amplitude on the record. As shown in Figure 7, the corratio of 2 on the record of the standard Woodrespondenceof peaks and troughsbetween the
Andersonseismograph.
two earthquakesis excellent,and there is no
The amplituderatiosof Love wavesfrom the difficulty in obtaining such ratios. Figure 8
two earthquakesobservedat Weston, Ottawa, showsthe ratio of the amplitude spectral denand ResoluteBay are shownin Figure 6. They sity obtained by the Fourier analysismethod.

SCALING

oO

o
,so,u'r.,|
WESTON

o7

LAW

OF SEISMIC

SPECTRUM

1225

samespectraldensityat short periodsfor the

two earthquakesand doesnot explain the ob-

I.

'Ms 0.85

servation.

A more general comparisonof the local

magnitudescaleM, and the surfacewavemag-

nitude scaleM, is difficult.The maximumamplitude recordedby the Wood-Andersonseismo-

//

3model

I0

20
PERIOD

IN

$0
SEC

graph would not be directly proportional to

the amplitude spectral density at any fixed
period,becausethe signaldurationand prevailing period may change with the earthquake
sourcesize. The spectral ratio may be nearly
equalto the maximumamplituderatio for such

Fig. 6. Comparisonof theoreticaland observed earthquakeswith small differencein magnitude

spectral ratio for two aftershocksof the Kern as studiedin the presentsection,but the equalCounty, California,earthquakeof 1952.Observed ity cannothold for larger magnitudedifference.
ratios are obtained from trace amplitude.

Further, the empiricalrelationbetweenM, and

M,, with which the theoreticalrelation is to be
compared,has not yet beenstabilized[Richter,

There is no significantdifferencebetweenthe
results obtainedby the two methods,justify- 1958].
ing the simpleprocedureusedby Berckhemer.
,

The theoretical curves of spectral density

RELATION BETWEEN ms AND M,
ratio for an earthquake pair with magnitudes
On the other hand, the relation between the
M, around 6.0 which best fit the observations
are shown in these figures. It is remarkable magnitude scale mB, defined as the logarithm
that the observedratio is about 7 at the period of amplitude of teleseismicbody waves, and
M, has been well establishedempirically by
of 20 sec; in other words,the differencein M,
between the two earthquakesis 0.85, about 3 Gutenbergand Richter [1965a]. Further, it is
timeslarger than the differencein M, obtained well known that the usual record of teleseismic
by Richter. Our u-squaremodel explainsthis body waves,obtainedby a standard short-pefact satisfactorily,becauseM, must have been riod seismographsuch as Benioff's, shows a
measuredon waves with periods of lessthan 1 rather narrow spectral band around I c/s.
sec, and this model predicts a spectral density Therefore, we may correlate the amplitude of
ratio of about 2 at these periods.On the other bodywaveswith the spectraldensityat I c/s.
If our signal is a finite portion of a Gaussian
hand, the u-cube model predicts nearly the
o)-

o)

b)./

IE
WBAY

RESOLUTE

IN
S

WESTON

MINUTE

S
IN

OTTAWA

Fig. 7. Love wavesfrom the Kern County aftershocks(no. 194 above,no. 141 below; num-

bersassigned
by Richter [1955]) recordedat Ottawa,Resolute,and Weston.

1226

KEIITI

o RESOLUTE
BAY
i-

WESTON

10

M, for the to-squareand .to-cubemodel. The

shaded area indicatesthe range between the

Ms----0.85

above-mentioned two extreme cases of the de-

n OTTAWA

pendenceof spectraldensityon signalduration.

M.=
Q3/,
/

The theoretical curve for the to-cube model does

'

,o4

-'

model
a modeI

....

I0
PERIOD

15
IN

20

AKI

25

SEC

Fig. 8. Comparison of theoretical and observed

spectral ratio for two aftershocks of the Kern
County, California, earthquake of 1952. Observed
spectral ratios are obtained by Fourier analysis.

noise, the amplitude spectral density will be

proportional to the square root of the signal
duration. On the other hand, if the signal is a
finite portion of a coherent sinusoidaloscillation, the spectral density will be proportional
to the signal duration. We may assumethat
the actual seismic signal has an intermediate
nature betweenthe above two extremes.Then,
we may write the spectral density at I e/s as

not agree with the empirical one given by

Gutenbergand Richter. On the other hand,the
agreementis excellentfor the to-squaremodel,
exceptfor smallermagnitudes.Looking at the
originaldata from whichGutenbergand Richter
derived their empirical formula, we find that
the theoretical curve based on the to-square
model better explainsthe observationsat small
magnitudesthan the empiricalcurve as shown
in Figure 10. This resultstronglysupportsthe

applicabilityof the scalelaw of seismicspectrum derived from the to-squaremodel on the

assumptionof similarity.
RELATIONBETWEENFAULTLENGTHANDM,
FORTHE (o-SQUAREMODEL

In derivingthe scalinglaw of seismicspectrum

we assumed that the characteristic

fre-

quency k is proportional to L -'. We can check

this assumption
againstgeological
or geodetic
observations
on an earthquakefault of known
magnitudeM,. The valueof k for a givenM,
is found from the theoretical curves for the

to-squaremodel shownin Figure 3. Then k-'

follows:

shouldbe proportional
to L, if the assumption

A(1) = c0nstX A= X t"z'

(34) of similarity holds.Figure 11 showsthe rela-

whereA., is the maximumtrace amplitudeand

t is the signalduration.
me
According
to GutenbergandRichter [1956b], 8
log t is relatedto m by the empiricalformula

log t = --1.9 -]- 0.4m

)-

(35)

Insertingthisequationinto (34), we obtain

log A(1) = cons-]- (1.2-, 1.4)ms

-cube

model

squa

(36)

From this equation and the charts of spectral

density curves given in Figures 3 and 4, we
can obtain the theoretical relation between m

and M, on the basis of the to-squareand to-cube

models.The constantin (36) is determinedin

such a way that m and M, agree at 6.75, in
accordance with the Gutenberg-Richter empirical formula

mB= 6.75 + 0.6a(M,- 6.75)

(37)
Figure 9 showsthe relationbetweenrns and

'ms=0,63Ms+ 2,5

(Gutenberg-Richter.

1956)

Ms

Fig. 9. Theoretical relation between ms and

M, based upon the -square and o-cubemodels,
as comparedwith the Gutenberg-Richterempirical formula.

SCALING

LAW

OF SEISMIC

SPECTRUM

1227

Ms-Ms

f[ [ I I I [ I I [ [ I ] i [ [
f,)- squar e moclel
Ms FROM
SURFACE
WAVES ,
1.21 0

' [ o

MMe =0.4(Ms-7)

.Sr
.4-

'

A KERN
COUNTY,
195Z

o o

ooo

2-

0-

0 AVERAGES
o

oo

oo

o o o o o

-.4-

o o

6-

8-

5.0

6.0

7.0

8.o

Ms

Fig. 10. Theoretical relation between ms and M, based upon the -square model, as compared with that observedby Gutenbergand Richter [1965a].

tion betweenTo -- 2,rk- and L for the earth

I0

2.0

To

50

sec

I00

200

500

I000

/
/

5OO

2OO
/

I00

oo / o o
/

,,/''"---L: CONST.
XTO
z

20
/

o/p/o

6.5 are excluded. In the case of small earth-

quakes,the surfaceevidencemay not revealthe

true fault lengthat the earthquakefocus.There
is alsosuchan ambiguitywith the magnitudeof
smallearthquakesthat the magnitudegiven in
Tocher'slist may or may not be taken as M,.
Consideringthesefacts,we may concludefrom
Figure 11 that geologicaldata do not exclude
the assumptionof similarity.
The characteristictime To for a given M, as
shownin Figure 11 may seema little too large.

It is possibleto reducethis value without affecting significantlythe conclusionsobtained

quake fault given in Toeher'slist [Tocher,

1960]. A linear relationshipholdsbetweenL
and To,if earthquakes
smallerthan magnitude

above. If we assume that k -- 10 vk. instead

of k = vkL,and if we determinea set of spectral density curves using the data of Berckhemer and others as given above,we find that

the value of To becomes about one-third that

o

t
6

6.5

I
7.5

8 --- Ms

given in Figure 11. The agreementbetween

theory and observationis as goodas that shown
in Figures5, 6, and.8, and we find again that
the to-squaremodel explains the relation be-

Fig. 11. Relationbetweenthe lengthof earthquake fault measuredby geologicalor geodetic tween ms and M, and that the to-cube model
means and the characteristic time of the earth-

quake determined from its magnitude on the

basis of the -square model. A linear relation
between them supportsthe assumptionof similarity.

does not.

EFFICIENCY OF SEISMIC RADIATION

Let us now examinethe efficiencyof seismic

1228

KEIITI

AKI

energy radiation, which must be independent tributed to a difference in the assumed source
of earthquakesourcesizeif the similaritycon- model. As mentioned before, the model of
dition holdsstrictly. We definethe efficiency Berckhemer,
if interpretedby dislocation
theory,
as the ratio of the energyradiatedin the form is the one in which the dislocation is constant
of seismicwavesto the elasticenergyreleased and independentof sourcesize,but the dislocaby the formationof an earthquakefault. If an tion in our modelis proportionalto the linear

earthquakeis a Starr fracture [Starr, 1928],

the fault-released
elasticenergyis proportional
to DolL. Under the assumption
of similarity,
this energywill be proportionalto L 3. If we
knowthe energyfor a certainvalueof Ms, we
can determinethe value for any M8 from the
scalinglaw givenin the preceding
section.Assumingthat log E is 23.7 for M, ---- 7.5 from
the resultof the writer'sstudyon the Niigata
earthquake[Aki, 1966], we get the released

dimension of the source.

DEPARTURE iROM SIMILARITY

As mentionedbefore,the assumptionof similarity implies a constant stress drop in all

earthquakes.If the stressdrop differs for two
earthquakes,our scalinglaw will not apply. If
the stressdrop varies systematicallywith respect to such environmental factors as focal
depth, orientation of fault plane, and cruststrain energyfor variousM, as shownin Table mantle structure, we may construct different
1. The energy radiated in the form of seismic scaling laws for different environments. Such
wavesis evaluatedby the Gutenberg-Richter a study of the seismicspectrummay eventually
formula
reveal the distributionof stressdrop or strength
of material in the earth's crust and mantle. For

log E = 11.4 + 1.5M,

(38)

and is also shownin Table I togetherwith its

ratio to the strain energy.The ratio definitely
increases with decreasingmagnitude. Thus,
starting with the assumptionof similarity, we
have endedby denyingit.
It is, however,not impossiblethat a further
refinementof the magnitude-energyrelation
(38) may eventually support the assumption
of similarity with regard to the radiation efficiency, because(38) is based upon several
simplifiedassumptions.

sucha study,however,we shall needmore precise measurements of spectrum over wider

rangesof frequency than are now available, as
well as detailed knowledgeof the propagation
factor of the spectrum.
Even with the present limited knowledgeof
the propagation factor, however, we may
demonstrate

remarkable

differences

in

stress

drops betweensome earthquakesby the use of

long-period surface waves. The earthquakesto
be comparedhere are the Niigata earthquake
of June 16, 1964, and the Parkfield (California)
earthquakeof June28, 1966.
It should be noted here that Bdth and Duda
The stress drop in the Niigata earthquake
[1964], usingBerckhemer's
result [Berckhemer,
was
obtainedby the following procedure[Aki,
1962], reached an entirely different conclusion
on the radiationeffciency.They found that the 1966]. The geometry of fault movement was
efficiencyincreaseswith increasingmagnitude determined from the radiation patterns of P
of the earthquake.This differencemay be at- waves,$ waves[Hirasawa,1966], and G waves.
The spectraldensity of displacementdue to G
waves was estimated for periods of 50 to 200
TABLE 1. ReleasedStrain Energy, Seismic
see, corrected for dissipation and geometric
Wave Energy and Efficiencyof
spreading,and comparedwith the theoretical
Seismic Radiation
excitation function [Haskell, 1964; Ben-Menahem and Harkrider, 1964] correspondingto a
log E,,t
log E,,,*
Ew[E,t
sourceof that geometry.From this comparison
8.5
26.7
24.2
0.003
we estimatedtheproduct of rigidity g, area $
8.0
25.2
23.4
0.016
of fault surface, and average dislocationAu,
7.5
23.7
22.7
0.10
which
correspondsto the moment Mo of the
7.0
22.5
21.9
0.25
component couple of the equivalent doublet
6.5
21.6
21.2
0.40
[Maruyama, 1963; Burridgeand Knopof],1964;
* log E,. = 11.4 q- 1.5 M..

Haskell, 1964]. The value of Mo (-- 1 AuS)

SCALING LAW OF SEISMIC

SPECTRUM

1229

for the Niigata earthquake was 3 X 10 dynes Parkfield earthquake. Consideringthat the efcm. All the near field evidence (echo-sound- feet of finite size was significantfor the Niigata
ing survey,aftershockepicenters,and Tsunami earthquake (about a factor of % at a period
source area) indicated a fault length, L, of of 70 see) but probably not for the Parkfield
about 100 kin. The focal depths of the main earthquake,we estimate the ratio of the source
shockand aftershocksindicated a fault width, momentMo for the Parkfield earthquaketo that
for the Niigata earthquake as 1/250. Thus, we
w, of about 20 kin. Assumingthat p -- 3.7 X
10 dynes cm-, correspondingto a shear ve- get a moment value of about 1 x 10 dynes
locity of 3.6 kin/see and densityof 2.85 g/cm, for the Parkfieldearthquake.
we obtainedthe value of the averagedislocation
Using the same rigidity value as for the
as 400 cm by inserting the valuesof L, w, and Niigata earthquake and the observedvalues of
p into the equationMo -- 1 Au Lw. This fault length and dislocationmentionedbefore,
value agreeswell with thoseobservedby echo- we get a fault width of about 13 km from the
soundingsurveys made just before and after above value of moment. This value of fault
the earthquake[Mogi et al., 1965]. Finally, the width gives us an extremely low estimate of
stress drop was estimated as about 125 bars strain release.Since Knopoff's fracture model
is more appropriate for a strike slip than
with the aid of Starr'stheory [Starr, 1928].
Now, let us comparethe Niigata earthquake Starr's, we estimate the strain releaseby the
with the Parkfield earthquake.The Parkfield formula e -- Au/2w [Knopo#, 1958]. We get
earthquaketook place right on the San Andreas a value of of 2 X 10-6 and a corresponding
fault

near

Cholame

and Parkfield.

The

PDE

stressdrop of about 0.7 bar, which is indeeda

card of the Coast and GeodeticSurvey reports

the epicenter as (35.9N, 120.5W), and the
origin time as 04:26:12.4 GCT, June 28, 1966.
The magnitudeis 5.8, 5.5, and 6 as givenby
the Pasadena,Berkeley, and Palisadesstations,
respectively.Accordingto a personalcommuni-

remarkably low value. Even if there is an order

of magnitudeerror in estimatingthe value of
moment,the stressdrop is still severalbars.
As mentionedbefore, if the stress drop. is
differentbetweentwo earthquakes,the scaling
law derivedin the presentpaper will not apply

cation from Clarence R. Allen and Stewart W.

to them. We found some indication of violation

Smithof the CaliforniaInstitute of Technology, of the scaling law when we compared the
the near field measurements revealed a strike
Parkfieldearthquakewith oneof the aftershocks
slip fault associatedwith this earthquake,its of the Kern Countyearthquake.
length being about 38 km and its offset about
The magnitude of the Parkfield earthquake
5 cm.

given by local stationsis 5.5 (Berkeley) , 5.8

G2 waves from this earthquake are clearly (Pasadena). The surface wave magnitude Ms
recordedby long-periodseismographs
at Reso- of this earthquake, calculated from the Love
lute (A = 40) and at Ottawa (A = 35). The wave amplitude at a period of 20 see recorded
peak-to-peak amplitudes on the records at a at Ottawa, is 6. This value agrees with the
period of 70 see are a little over 1 mm at both magnitudegivenby the Palisadesstation.
On the other hand, the magnitude of numstations.This correspondsto a spectraldensity
ber 141 aftershock[Richter, 1955] of the Kern
of ground displacementof about 0.04 em seeat
County earthquake is 6.1. Ms for this earththat period.
The G2 waves from the Niigata earthquake quake, calculatedalso from Love wave amplitude at a period of 20 seerecordedat Ottawa,
at, the epicentraldistanceof 35 to 40 showa
spectraldensityof about 1.6 em seeat a period is 6.2.
of 70 see for a certain radiation azimuth. If the
Since the variability of seismicamplitudes is
Niigata earthquakesourceis a strike slip fault very large, it is dangerousto draw any conclulike the Parkfield earthquake, and if we ob- sions from measurements at a few stations.
served G waves in the direction of maximum
However, the magnitudevalues above suggest
radiation,we wouldexpecta spectraldensityof that the spectral density for the Parkfield
about 5 em see at a period of 70 see for G2 earthquake may be greater than that for the
waves at A -- 35 40 . This value is about
Kern County aftershock at long periods, and
125 times as large as that observedfrom the smallerat short periods.If so, the two spec-

1230

KEIITI

AKI

trum curvesmust crosseachother, violating the Berckhemer,H., Die Ausdehnungder Bruchfiiche

i'm Erdbebenherd
und ihr Einfiussauf dasseisscalinglaw. This result is expectedif the stress
mische Wellenspektrum, Getlands Beitr. Geodrop in the Parkfieldearthquakeis lower than
phys.,71, 5-26, 1962.
that in the Kern County aftershock.The reduc- Burridge, R., and L. Knopoff, Body force equivation of stressdrop is equivalentto the reduction
lents. for seismic dislocations, Bull. $eismol.
$oc. Am., 54, 1875-1888,1964.
of Do in (30), and it will shift the spectrum
curves in Figure 3 downward parallel to the Byefly, P., The periodsof local earthquakewaves

ordinate,causingan intersection
with the original curve in the manner described above.

in central California, Bull. $eismol. $oc. Am.,

37, 291-298, 1947.

Gutenberg, B., and C. F. Richter, On seismic

As we haveseenabove,there is a possibility waves, Getlands Beitr. Geophys., 47, 73-131,

that the stressdrop in an earthquakemay vary

1936.

greatly accordingto its geologicalenvironment. Gutenberg,B., and C. F. Richter,Earthquake

magnitude, intensity, energy, and acceleration,
We shall probablyhave to assigndifferentscalBull. $eismol. $oc. Am., 32, 163-191, 1942.
ing lawsto differentenvironments.
This implies Gutenberg, B., and C. F. Richter, Earthquake
that a singleparameter,such as magnitude, magnitude, intensity, energy, and acceleration,
2, Bull $eismol. $oc. Am., 46, 105-145, 1956a.
cannotdescribean earthquakeevenas a rough
measure.The measurementof seismicspectral Gutenberg, B., and C. F. Richter, Magnitude and
energy of earthquakes, Ann. Geofs. Rome, 9,
density rather than amplitude will becomeincreasingly important. To understand the ob-

1-15, 1956b.
Haskell, N., Total energy and energy spectral
density of elastic wave radiation from propagating faults, Bull. $eismol. $oc. Am., 54, 1811-

servedspectrumin terms of the physicsof the

earthquakesource,however,we shall have to
knowmoreaboutthe effectof the propagation 1842, 1964.
Haskell, N., Total energy and energy spectral
mediaon the spectrumthan we do now.
density of elastic wave radiation from propagating faults, 2, A statistical sourcemodel, Bull.
$eismol. $oc. Am., 56, 125-140, 1966.
lute recordsof the Parkfieldearthquakeavailable Hirasawa, T., Source mechanismof the Niigata
to me.
earthquake of June 16, 1964, as derived from
This researchwas supportedby the Advanced
analysisof body waves, J. Phys. Earth, 14, in

Acknowledgments. I should like to thank Dr.

J. H. Hodgson for making the Ottawa and Reso-

ResearchProjectsAgencyand was monitoredby

the Air Force Office of Scientific Research under
contract AF49 (638)-1632.
EFERENCES

Aki, K., Corre]ogram

analysisof seismograms
by
meansof a simpleautomaticcomputer,J. hys.
Earth, 4, 71-79, 1956.

Aki, K., Generationand propagationof G waves

from the Niigata earthquakeof June 16, 1964,

2, Estimation of earthquakemoment, released
energy, and stress-straindrop from the G wave
spectrum,Bull. Earthquake Res. Inst. Tokyo
Univ., 44, 73-88, 1966.

Asada,T., On the relation betweenthe predominant period and maximum amplitude of earth-

quake motions,J. $eismol.$oc. Japan,$er. 2,

6, 69-73, 1953.

Bth, M., and S. J. Duda, Earthquakevolume,

fault plane area, seismicenergy, strain, deformation, and related quantities, Ann. Geofis.
Rome, 17, 353-368, 1964.
Bath, M., and C. F. Richter, Mechanism of the

aftershocksof the Kern County, California,

earthquakeof 1952,Bull. $eismol.$oc. Am., 48,
133-146, 1958.

Ben-Menaham, A., and D. G. Itarkrider, Radiation patterns of seismic surface waves from
buffed dipolar point sources in a fiat stafffled
earth, J. Geophys.Res., 69, 2605-2620,1964.

press,1966.

Honda, H., and H. Ito, On the period of the P

waves and the magnitude of the earthquake,
Geophys.Mag., 13, 155-160, 1939.
Jones, A. E., Empirical studies of some of the
seismic phenomena of Hawaii, Bull. $eismol.
Soc.Am., 28, 313-338, 1938.
Kanai, K., K. Osada,and S. Yoshizawa,The relation between the amplitude and the period
of earthquake motion, Bull. Earthquake Res.
Inst. Tokyo Univ., 31, 45-56, 1953.
Kashara, K., The nature of seismicorigin as inferred from seismologicaland geodetic observations, 1, Bull. Earthquake Res. Inst. Tokyo,
35, 747-532, 1957.
Knopoff, L., Energy releasein earthquakes,Geophys. J., 1, 44-52, 1958.
Maruyama, T., On the force equivalentsof dynamic elastic dislocations with reference to the

earthquake mechanism,Bull. Earthquake Res.

Inst. Tokyo Univ., 41, 467-486, 1963.
Matumoto, T., On the spectral structure of earth-

quakewaves,Bull. EarthquakeRes.Inst. Tokyo,

38, 13-28, 1960.

Matuzawa,T., Study of earthquakes,OhO$hoten,

Tokyo, pp. 45, 207, 1964.

Mogi, A., B. Kawamura, and Y. Iwabuchi,Submarine crustal movement due to the Niigata
earthquake in 1964, in the environs of the Awa

SCALING LAW OF SEISMIC

Shima Island, Japan Sea, J. Geodetic. oc.

Japan, 10, 180-186, 1965.
Press,F., Long period waves and free oscillations

of the earth, in Researchin Geophysics,vol. 2,

chapter 1, pp. 1-26, The M.I.T. Press, Cambridge, Mass., 1964.
Richter, C. F., Foreshocksand aftershocks,Earthquakesin Kern County, California, during 1952,
Calif Dept. Nat. Resources,Div. Mines, Bull.

SPECTRUM

Starr, A. T., Slip in a crystal and rupture in a

solid due to shear, Proc Cambridge Phil. Soc.,
24, 489-500, 1928.

Tocher, D., Movement on faults, Proc. 2nd World

Conf. Earthquake Engineering, 1, 551-564, 1960.
Tsuboi, C., Earthquake energy, earthquake volume, aftershock area, and strength of the earth's
crust, J. Phys. Earth, 4, 63-66, 1956.

171, 177-198, 1955.

Richter, C. F., Elementary Seismology,p. 347,

W. tI. Freeman and Co., San Francisco,1958.

1231

(Received September 17, 1966.)