University of Nebraska - Lincoln
DigitalCommons@University of Nebraska - Lincoln
USGS Staff -- Published Research
US Geological Survey
1-1-1973
Seismicity Preceding Moderate Earthquakes in
California
Robert L. Wesson
U.S. Geological Survey
William L. Ellsworth
U.S. Geological Survey
Follow this and additional works at: http://digitalcommons.unl.edu/usgsstaffpub
Part of the Earth Sciences Commons
Wesson, Robert L. and Ellsworth, William L., "Seismicity Preceding Moderate Earthquakes in California" (1973). USGS Staff -Published Research. Paper 382.
http://digitalcommons.unl.edu/usgsstaffpub/382
This Article is brought to you for free and open access by the US Geological Survey at DigitalCommons@University of Nebraska - Lincoln. It has been
accepted for inclusion in USGS Staff -- Published Research by an authorized administrator of DigitalCommons@University of Nebraska - Lincoln.
VOL. 78, NO. 35
JOURNAL OF GEOPHYSICAL RESEARCH
DECEMBER 10, 1973
Seismicity
Preceding
Moderate
Earthquakes
in California
ROBERT L. WESSO• A•
WILLIAM, L. ELLSWORTH
National Center ]or Earthquake Research,U.S. GeologicalSurvey
Menlo Park, ,California 94025
Regional seismicitydata have been examinedto determinethe nature of seismicityin the
yearsprecedingthesemoderateearthquakes
in California: 1952Kern County (M '- 7.7),
1963Watsonville (M -- 5.4), 1964and 1967Corralitos (M -- 5.0, 5.3), 1966Parkfield-Cholame
(M '-- 5.1, 5.5), 1968BorregoMountain (M = 6.4), 1969Santa Rosa (M '= 5.6, 5.7), and
1971San Fernando (M -- 6.4). In each instancethe moderateearthquakesoccurredin areas
characterized
by a relatively high level of small earthquakeactivity. In most casesthe pre-
cedingactivityin the areaimmediatelysurrounding
the epicenterwashighrelativeto o•her
segments
of the samefault zoneor otherneaxbyfault zonesthat, on the basisof geologic
criteria,would be considered
equally likely locationsfor moderateearthquakes.
Detailed
observationsfor a moderateearthquakethat occurredin 1972near Bear Valley (M _-- 5.0),
where seismograph
station coverageis adequateto obtain reliable focal depthsfor small
earthquakes,
indicatethat concentrations
of smallearthquakes
werepresentin the immediate
hypocentralregionin the monthsbeforethe earthquake.The regionalobservations
have
potentialapplications
for the a•alysisof seismic
risk in that fault zonesor segments
of fault
zonescharacterizedby a relatively large number of small earthquakesseem more likely to
sustainmoderateearthquakesthan do adjacentfault zonesor fault segmentswith equivalent
geologicevidencefor recencyof movement.The detailedobservations
providea potential
tool for mappingstressconcentrations
or areasof anomalouslylow strengthalongthe parts
of fault zonescapableof seismicslip.
The relation between small earthquake activ-
it has been demonstrated that most small earth-
in Californiaare closdy
ity and the potential for moderate or larger quakesand aftershocks
earthquakespromisesto providepowerfultools associatedwith geologicallymappable faults
for the predictionof earthquakes.(For purposes [Lee, 1971; Eaton et al., 1970a, b; Hamilton,
of this discussion, 'small earthquakes' means 1972; Wessonand Wilson, 1972]. In fact, it
earthquakes with local magnitudes less than has becomepossibleto 'map' many faults by
magnitude5. In this term we include'micro- the distribution of small earthquakes along
earthquakes,'which are generallytaken to be them. In contrast, several sections(as long as
earthquakeswith local magnitudeslessthan 3.) 300 km) of major fault zonesthat have sus-
tainedlarge displacements
in historictime display a very low level of seismicactivity at
damaging shallow earthquakes occur along present [Lee, 1971; Brune and Allen, 1967;
faults that can be (or couldhave been) demon- Page, 1972; Allen et al., 1965; Ryall et al.,
strated by geologicmeansto have sustained 1972; Byefly, 1937]. If studiesof small earthdisplacement
in Quaternarytime [e.g., Went- quakes are to be useful in earthquakepreworth and Yerkes, 1971; Brown et al., 1967; diction, we must discoverhow the presenceor
Allen et d., 1968; Clark et al., 1972; Dibblee, absenceof small earthquake activity rela.testo
1955]. Before the developmentof denselocal the potential for a moderateor large earthseismographnetworksin the 1960'sthe accu- quake.
The answer to this question has in the past
racy of locatingsmallearthquakeswas suchas
to suggestthat they occurredrather generally been shroudedby the difficultiesof accurately
and by uncertainty
throughoutan active region.With the devel- locatingsmallearthquakes
opmentof denseseismograph
networksand with aboutthe time scaleon whichprecedingactivity
careful attention to crustal velocity structure is important.Exceptfor somewall-defined
foreExtensive evidence has been accumulated in
the western United States and elsewhere that
Copyright¸
shocks that occur within hours prior to a
1973 by the American GeophysicalUnion.
8527
This article is a U.S. government work, and is not subject to copyright in the United States.
8528
WESSONAND ELLSWORTH.' SEISMICITY iPRECEDING
EARTHQUAKES
moderateor large earthquake,there is little or
no evidenceto suggesta buildupof smallearth-
ship--for example, between the focal mechanisms and source parameters of the preceding
quakesin the daysandweekspriorto themain small earthquakesand those of the main shock.
shock [Richter, 1955a; Allen, 1971; McEvilly
Unfortunately, the data available offer only a
et al., 1967]. Richter has observed, however, crude suggestionof a spatial relationship; dethat 'In 28 years of seismograph
recordingfrom scription of a causal relationship must await
1929 through 1956 there was only one earth- further studies.
quake in southern California of magnitude 5, or
To comparethe pattern of seismicityprior to
over, which was not precededby others from the main shock with the rupture surface or
the same or an adjacent epicenter, although surfaces,we have characterizedthe main shock
weeksor monthsmight intervenebetweenthe by its zone of aftershockactivity, as has been
last previous shockand the large one [Richter, done by Fedotov [1965], Mogi [1968a], and
1958, p. 67].'
Sykes [1971]. This characterization has two
We have examined the seismicity data for qualities to recommendit. First, the zone of
some regions in which moderate earthquakes aftershocksseemsto be a good approximation
have occurredin California, for which adequate of the area of slip during the main event. Secinstrumental capability existed prior to the ond, if the earthquakesused to define the aftermain shock, in order to determine the nature shock zone have been located in the same way
and distribution of the small earthquakes, if
as thoseprecedingthe main shock,the definition
any, that occurredin the months and years of the aftershock zone contains an estimate of
precedingthe main shock. Similar work has the uncertainty of the locations.For this reason
been done in Japan by Mogi [1968b, 1969] we have generallychosenaftershockzonesas
and in the Soviet Union by Keylis-Borok and defined by the earthquake locations from reMalinovskaya [1964] and Fedotov [1968].
gionalnetworks,
ratherthan the moreprecise
definition
sometimes available
from
studies in
ME?HOD
which many additionalstationsare deployed.
The events selected for study are listed in
To study the seismicityof regions of California in which moderate earthquakes have Table 1. The distanceto the closests•ation (or
occurred and for which an adequate instrumen- stations) is includedto give someindicationof
tal capability existedprior to the earthquakes, the detection and location capability prior to
we examinedthe publishedbulletins (and some the main shock. The locations of these events
asyet unpublished
data) fromthe Seismologicalrelative to the major faults of California are
Laboratory,CaliforniaInstitute of Technology indicated in Figure 1.
In each of the moderate earthquakesstudied
(Caltech), and from the Seismographic
Stations,
University of California, Berkeley (UCB), for we have selecteda period for which we believed
the periodsince1962.In additionwe examined the locations were most reliable, or were at
the Caltech data for the twenty years prior to least consistent.We have plotted the epicenters
the 1952 Kern County earthquake. Both seis- of earthquakeswithin a regionwith dimen•ons
mographnetworksoperatedby these two institutionsare regionalin nature, and both grew
significantlyduring the 1960'•. Both groups
attempt to make their lists of earthquakes
substantiallycompleteabovemagnitude3, with
many smaller earthquakeslocated as well. Under favorable circumstances the accuracy of
a few to several times that
of the aftershock
zone.
OBSERVATIONS
1952 Kern County earthquake. The Kern
County earthquake(M -- 7.7) of July 21, 1952,
is the largestearthquakerecordedin California
epicenter location is probably within a few
kilometers; under unfavorable circumstances
the error is doubtlesssomewhatlarger.
To make a truly convincingcasethat small
and microearthquake activity is premonitory
to a moderateor large earthquake,it would be
since a sufficient number of stations existed to
desirable
[Dibblee, 1955; Buwalda and St. Armand,
to
establish
a mechanistic
relation-
detect and locate preceding activity. In this
earthquake,
thrustingin a northwesterly
direction accompaniedby left lateral slip occurred
alonga 50-km sectionof the White Wolf fault
at the southern end of the San Joaquin Valley
WrssoNANDELLSWORTH'
SEISMICITY
PRECEDING
]•ARTHQUAKES
o
o
O•
o
0
o
•
0•-•
o
•-•
•-•
0
O0
•0
00•
•
8529
8530
WESSON
ANDELLSWORTH;
SEISMICITY
PRECEDING
EARTHQUAKES
SANTA
ROSA
Oct. 2,
CORRALITOS
8,1967
16, i964
WATSONVILLE
BEAR
Sept. 14, I
VALLEY
24,1972
PARKFIELD
--
dune 28, 1966
KERN
June
SAN
FERNANDO
Feb. 9,1971
0
I00
COUNTY
21, 1952
•
200KM BORREGOMOUNTAI
April
9,
1968
Fig. 1. Map of California
showing
the outlines
of the sevenregions
surrounding
the earthquakes.
Alsoshownarethe majorfaultsof theSanAndreas
faultsystem.
1955]. Geologicevidencepoints to more than
3 km of verticaloffsetalongthis fault sincethe
beginningof Pleistocene
time. However,the
lock faults,both of whichwerevery activein
Quaternarytime. The regionof the SaNAndreas
fault south an.d west of the White Wolf fault
fault hadnot beenrecognized
as a contempo-wasthe site of a great earthquakein 1857,with
rary earthquakehazardprior to the 1952earth-
ground breakage that extended for 250 km or
quakes[Oakeshott,
1955].Richter[1955b],in more [Richter, 1958]. Of particular interestin
discussing
activity preceding
the earthquake, the 1952 earthquakeis that, although the
concluded,'There was little prelude to the regionof the aftershockzone containingthe
major earthquake
of July 21, 1952.'Although epicenterof the main shock,and presumably
the data availablesuggestthat this is true on the initiation of rupture, was relativelyactive
time scalesof lessthan a few years,it is cer- beforethe main shock,the oppositeend of the
tainlynot true on a time scaleof manyyears. aftershockzonewas relatively quiet. The afterLocations
of earthquakes
in the 20 yearspre- shockswere concentratedat oppositeends of
cedingthe main shock[Richter,1955a] show the zone.
that the activity within about 15 km of the
1963 Watsonville and 196•, and 1967 Cottaepicenter
wassubstantially
higherthanactivity lifos earthquakes. Earthquakesin the Watsonin mostof the surrounding
region(seeFigure ville-Corralitosarea are of particular interest
2). In particular,the activityin the vicinityof becausethey are locatedat the northern termithe epicenterwas very muchgreaterthan the nus of the 200-km-longzoneof relativelyhigh
activity along either the San Andreas or Gar-
seismicityalong the San Andreas fault in cen-
8531
WESSON AND ELLSWORTH' SEISMICITY PRECEDINGEART}IQUAKES
,
ß
ß
35o45 '
.1• ;32 to
ß •. I/I
ß
ß •• 2.. .
BAKERSFIELD
• •
/
/ •
I
.
e'
..
'
7/21/52
MAGNITUDE
• ß
•3.0
ß
3.0 - 3.9
ß
4.0
ß
4.9
MAIN
SHOCK
(•,
.
ZONE OF
AFTERSHOCKS
ß
35ø00 '
ß ee
ß
ß ß
PINE
ß
ß
ß
34'30'
119o00 '
120'00'
118'00'
Fig. 2. Seismicityof the Bakersfieldregionfrom 1932to the occurrence
of the M = 7.7
Kern Countyearthquakeof July 21, 1952,as reportedby Richter [1955a].Clustersof solid
circlesrepresentgroupsof epicenterswith the samereportedlocation.The epicenterof the
main shock,markedby a hexagon,a•d the aftershockzone,enclosed
by the dashedlines,are
from Richter
[1955b].
tral California [Bolt et al., 1968; Bolt and
Miller, 1971]. Northward from this area to
Cape Mendocinothe San Andreasfault displays
a relatively low level of seismicity.The 35-km
segmentof the San Andreas system south of
Corralitos was one of the most active, generating 7 earthquakeswith magnitudesgreater than
4.0 between
1963 and 1967.
Three of these earthquakes (M -- 4.4, 4.2,
5.4) occurredalong the San Andreasfault zone
in the vicinity of Watsonville during the late
During the 3 yearsprior to the initiationof
thesemoderateearthquakesthis area displayed
a relatively high level of seismicity. Ud•as
[1965, 1969] reportedan averageof 2-3 earth-
quakes per month in the magnituderange
2.5-3.5 in the vicinity of Watsonville.The seis-
micity in this areafor the 16 monthspreceding
the first moderateearthquake (M = 4.4, May
7, 1963) was muchgreaterthan the seismicity
either to the northwest or in the adjoining segment of the San Andreas fault zone to the
springand summerof 1963.Althoughno ground southeast,directlyeast of Salinas(Figure 3a).
breakagewas reportedfor theseeventsand the This relationshipof a relatively active zone
instrumental locations indicate epicenters of
about 5-10
km to the southwest of the fault
zone, the earthquakesprobably occurredmuch
closer,perhaps along the fault itself, as more
recent work
on crustal velocities and station
delay suggests[McEvilly, 19'66; Wessonet al.,
1•?•].
adjoined
oneitherendby a quietzonepersisted
throughoutthe sequence,
althoughthe active
zoneexpandednorthward.The large relative
and absoluteerrors in the epicentrallocations,
suggested
by the subsequent
work on small
earthquakes
in the areaby Leeet al. [1972a,b,
el, makeit difficultto draw firm conclusions
8532
WESSONAND ELLSWORTI-I:SEISMICITY PRECEDINGEARTHQUAKES
Fig. 3. Seismicity of the Watsonville-Corralitos area for the periods indicated as reported
by UCB. The aftershock zone of the M = 5.4 earthquakes of September 14, 1963, is after
Ud•½s [1965]. The aftershock zone of the M = 5.0 Corralitos earthquake of November 16,
1964, is after McEvilly [1966]. No aftershocks were reported for the M -- 5.3 Corralitos
earthquake of September 7, 1967.
about the relationbetweenthe seismicityprecedingthe moderate earthquakesand the moderate earthquakesthemselves.However, it is
clear (1) that the M = 4.4 earthquakeof May
7' occurred at the northern end of the active
zone, (2) that the M = 4.2 earthquake of
August 31 occurred toward the southern end
of the active zone, (3) that the M = 5.4 earth-
quakeof September14 occurredadjacentto the
epicenterof the May 7 earthquake,and (4) that
the aftershocksof the M = 5.4 earthquake extended the zone of activity northwestward(Figure 3b).
After the M = 5.4 earthquakeon September
14, 1963, and its aftershocks,activity increased
in the vicinity of the Sargent fault, culminating
in an earthquakeof M = 5.0 on November 16,
1964 (McEvilly [1966]; seeFigure 3c). In the
18 months prior to this earthquake, 3 earthquakeswith magnitudesgreater than or equal
WESSONAND ELLSWORTH:SEISMICITY PRECEDINGEARTI-IqUAKES
to 3.0 occurred within
5 km of the eventual
epicenter (Figure 3b, c). After the M -- 5.0
earthquake,activity beganto increasealongthe
San Andreas fault immediately to the west of
the epicenter(Figure3d), includingearthquakes
of magnitudes4.2 and 4.7 (October 14, 1966,
and September 7, 1967). This activity contrasted markedly with the quiescenceof the
area in the precedingthree years (Figure 3a-•).
On December18, 1967, a M -- 5.3 earthquake
occurredin this area (Figure 3d). Relocationof
the epicentersof the September7, 1967, earthquake and the main shockby usingthe station
8533
quiescentin the months precedingthe earthquake. McEvilly et al. [1967] conclude that
the 'first half of the year 1966 was characterized
by apparently normal or •ightly increasedlevel
of seismicityin the general Parkfield area.'
The 1966 preearthquakeactivity clustersjust
to the north of the epicenterof the main shock.
Althoughthe 1966preearthquakeactivity shown
in Figure 4 is not a complete set over the
entire map area, T. V. McEvilly (personal
communication, 1973) believes that it representsa completepicture to magnitude2.5 along
the
San
Andreas
fault
between
Priest
and
correctionsof Lee et al. [1972a] indicates that
the epicentersof thesetwo eventswerelessthan
2 km apart and that they both occurred at
Cholameand therefore that the apparent clustering to the north of the epicenter of the
main shock is significant.In other words, the
depths between 10 and 12 km. No surface activity just north of the epicenter was signifibreakageor aftershockswere observedafter the cantly greater than that either in the afterDecember 18 earthquake, and therefore no con- shock zone to the south or along the fault
venient estimateof the rupture surfaceis avail- farther north in the 6 months prior to the main
able, but a sourcedimensionof a few kilometers shocks.
can be inferred from the moment calculated for
1963 Borrego Mountain earthquake. The
the earthquakeby Wyssand Brune [1971] and San Jacinto-Imperial fault zone, which splays
from the magnitudeby usingscalingarguments off from the San Andreas fault north of San
[Dieterich, 1973].
1966 Parkfield-Cholame earthquakes. The
Parkfield-Cholamc earthquakesof June 28, 1966
(M -- 5.1, 5.5), occurredin an area that has
been the site of several earthquakesin the magnitude 5-6 range in this century [McEvilly et
a/., 1967]. Their occurrencewas therefore not
surprising on a time scale of tens of years.
McEvilly et al. [1967], Allen and Smith [1966],
and Eaton [1967] have discussedearthquake
activity in the months prior to June 28. The
area is on the periphery of the UCB reporting
area, and therefore the routine detection and
location of small shocksis not as complete or
accurate as might be desired. The estimated
uncertainty in the locations for the 1966 preearthquakeactivity (Figure 4) is 2-5 km north
of Parkfield
to 5-10 km near Cholame.
Although only 11 earthquakeswere located
in 1966 preceding the earthquake, 43 events
with S-P intervals less than 4 see (epicentral
distanceslessthan about 30 km) were detected
by the seismographstation at Gold Hill during
the same period of time [Eaton, 1967]. There-
fore, although it has been reported that no
earthquakeswere recordedin a 24-hour period
during June [Allen and Smith, 1966], it is not
true
that
the
Parkfield-Cholame
area
was
Bernadino
and extends southward
to the Mexi-
can border and beyond, has been one of the
most active areas in California during this
century [Allen et al., 1965]. Nine earthquakes
of M -- 6.0 or greater have occurredalong this
300-km-long zone of subparallel faults since
1915, includingthe M -- 7.1 E1 Ccntro earthquake of 1940. Ground breakage of the M -6.4 Borrcgo Mountain earthquake of April 9,
1968, occurred along a 33-km segment of the
Coyote Creek fault, which lies within the San
Jacinto Fault zone [Allen et al., 1968]. Aftershocksof the earthquake located by Hamilton
[1972] definea zoneof activity stretchingabout
50 km along the fault zone and two clusters,
one 20 km northeast of the epicenter and one
20 km southwestof the epicenter.The breadth
of the primary aftershockzone (in contrast to
that of the Parkfield earthquake [Eaton et al.,
1970b]), the observedslip on the Superstition
Hills fault and on the San Andreas fault at the
time of the earthquake [Allen eta/., 1968], and
the apparent aftershockactivity in spatially
separatedregionsall suggesta more complex
nature for strain release in this area than
the Parkfield area.
for
The epicenterof the 1968 BorrcgoMountain
earthquakelies 30 km northeast of the M --
8534
WEssoN ANDELLSWORTH:SEISMICITYPRECEDING
EARTHQUAKES
36'30'
oO
o
0
o
0
0
_
000
0 '
_
•
aftershocks•
MAG. to
to
6-28-66
•.•-3.o o
3.1-4.0 0
ß
I
•
Q
4.,-4.70
.
•_
1965
6-27-66
o
•.•
Parkfield
• •
Cholame
••/
6-
•
•_•,_•
•
Seismogr.ph
st.
tion
•• 0 .
25KM
c",
I
•
m
•bO'
•
•
•
35'30'
•0'00'
Yi•. 4. •ei•miciW of •he P•rkfield-•hol•me •e•ion recorded before •he
: 5.• e•r•hqu•ke• oL June •, 1•6. Open circle• repre•en• epicen•er•from •he
c•!b• for1• •hrou•h
•.
•oc•fio•reported
b?••
e• •. [1967]
for1•6 appear
a• •olid circle•. Two •hock• recordedon June •, 1•, before •he m•in •hock• •re •hown
•qu•re•, •nd the epicen•er•of •he moderatee•r•hqu•ke• •ppear • hexagons.
6.5 shockof October21, 1942, and 10 km southeast of the M ---- 6.2 earthquake of March
1954. Although the slip surfacesof these two
earthquakesare not known in detail, Allen et
Comparisonof seismicactivity in the Borrego
Mountain
region
forthe5 years
preceding
April
1968 (Figure 5) with other active areas in
California clearly demonstrates
the extremely
al. •1968]statethat 'no knownmajorhistoric high level of activity over most of the San
activity had occurredalong the exact segment
broken on April 9th.' Examination of records
from a Caltech temporary station operatedat
ObsidianButte, 47 km from the epicenter,by
Allen and Nordquist [1972] led them to coneludethat no detectableactivity occurredin the
days precedingthe event. They also reported
that the level of activity at the ObsidianButte
station during the four monthsprecedingthe
earthquakewas at a very low level.
Jacinto fault zone. Aftershocks of the M = 4.7
earthquake of May 21, 1967, and adjoining
epicentersfrom the period 1963-1965 define an
active zone 30 km to th• northwest of the
1968 aftershock zone along the San Jacinto
fault zone.A secondzoneof high activity lies to
thesouthof theSaltonSeaalongan extension
of
the ground breakage of the 1940 E1 Centro
earthquake. Most of these events occurred in a
WESSON AND ELLSWORTH' SEISMICITY PRECEDINGEARTHQUAKES
i
ß
55'45'
t
i
8535
ß
1-65
to
MAG.
7-65
to
6-65
•= 5.0
o
5.1 - 4.0
o
4-68
ß
4.1 - 5.0
ZONE
OF
AFTERSHOGKS
Main
shock
•o
ß
0 oO
o
o
o
o
El Gentro
ß
GALtFORNtA
U.S.A.
................
........................
25
o
ß
'--"f
KM
o
ß
I
I
116ß45 '
I
I
52*50'
115'50'
Fig. 5. Seismicity of the Borrego Mountain region from the Caltech catalog prior to the
M -- 6.4 Borrego Mountain earthquake of April 9, 1968. The epicenter of the main shock
and the aftershockzone are after Allen and Nordquist [1972].
swarm in June 1965 that included 5 events
greaterthan magnitude4.
The cluster of earthquakes adjoining the
1968 aftershock zone to the east occurred in
late August 1965. The Caliente fault zone to
the west of the aftershock zone showedan intensification of activity late in 1966 that continued through 1967.
The immediate epicentral region of the
complexinterconnectionof faults, the activity
along the subparallelSan Jacinto fault may be
a result of the same phenomena observed to
precedeother earthquakesin this study.
1969 Santa Rosa earthquakes. Two earthquakes of M -- 5.6, 5.7 shook the city of Santa
Rosa on October 2, 1969. McEvilly [1970]
characterized the area around Santa Rosa as
having had 'a fairly active seismic history.'
Borrego Mountain parthquakeis not charac- First motions and the distribution of afterterized by a high level of activity during the
years preceding the earthquake, although it
adjoinsthe clustersof activity recordedin August 1965. However, given the complexity of
the geologicstructure in this area and the
shocksdeterminedby Unger and Eaton [1970]
show right lateral slip on a northwest-trending
fault that runs subparallelto the Healdsburg
fault to the north. Locations for the earthquakesreportedby UCB for the 8 yearspreced-
WESSON AND ]•LLSWORTH' SEISMICITY PRECEDING]•ARTHQUA.
KES
8536
1971 San Fernando earthquake. The San
ing October2, 1969 (Figure 6), are scattered,
particularlyfor the smallerevents,but clearly Fernando earthquake (M -- 6.4) of February
a relatively high level of activity existed in 9, 1971, occurredon faults not previouslyrecogthe immediate epicentral area. Of particular nized as active, although ample geologiceviinterest is the M -- 4.6 earthquake that oc-
curred on April 25, 1968, with an epicenter
nearly identicalto that of the marinshocks.No
earthquakegreater than magnitude4 occurred
in the vicinity of Santa Rosa for at least the
previous25 years [McEvilly, 1970]. The activity in the vicinityof SantaRosais isolatedand
representsa level of seismicitysubstantially
greater than that associatedwith the San
Andreas fault to the west, along which ground
breakageoccurredwith the 1906 earthquake,
or with the Healdsburg-RodgersCreek fault
zone either to the northwest or southeast of
Santa Rosa.
dence existed before the earthquake to demonstrate movement along their traces in late
Quaternary time [Wentworth and Yerkes,
1971]. The earthquake was accompaniedby
thrusting and left lateral slip along a westwardtrending, northward-dipping zone of thrust
faults
now
termed
the
San
Fernando
fault
zone. Prior to February 1971, no moderate or
large earthquake had occurred in this area
since instrumental epicentershave been determined. Allen [1971] discusseda large damaging
earthquake that occurredsomewherein the area
in 1893. In trenching fault scarps from the
earthquake, Bonilla [1973] found wood frag39o00 '
ß
ZONE OF
AFTERSHOCKS
ß
M 5.6
M 5.7
Sebastapol
Napa
I/I/62
to
10/2/69
Petaluma
MAGNITUDE
ß
2.5
ß
5.1
ß
-
- 4.0
>4.1
Main
0
5.0
shock
15 KM
123015 '
38000 '
122015'
Fig. 6. Seismicity of the Santa Rosa area prior to the M -- 5.6 and M -- 5.7 Santa Rosa
earthquakesof October 2, 1969, from the UCB catalog and McEvilly [1970]. The aftershock
zone is after Unger and Eaton [1970].
WESSONANDELLSWORTI-I'SErSMrCrTYPRECEDING
EART•QrJAI<ES
8537
119' 15'
$5'15'
\,
Aftershock
Zone of /
1952
Kern Co
/
eorfh½
/
I
MOJAVE
DESERT
\
Santa
L
_
i,;I
f(•ult
•• LOS
ANGEL•
PACIFIC
OCEAN
0 • 15
I
I
•
•
'.••!933L.g
Beach
•
__
• ••
•%
•
33' 45'
117' 45'
Fig. 7a.
Fig. 7. (a) Major faults, cultural features, and aftershock zones, and (b) seismicity preceding the M = 6.4 San Fernando earthquake of February 9, 1971, from the Calteeh catalog.
The aftershock zone is after Allen et al. [1971] and Wesson et al. [1972]. The location of
the main shock is after Allen et al. [1971].
mentsin what he interpretedas collapserubble in the northernmostarea representcontinuing
from a scarp resultingfrom a previousearth- activity in the aftershock zone of the 1952
quake.Dating by the carbon14 methodindi- Kern County earthquake.Earthquakesin the
cared an age of about 200 years, suggesting southernmost
area occurovera wideareasouth
that an earthquake causingsurface rupture of the Santa Monica mountains.The San Feroccurredsometimein the 1700's,perhaps the nando earthquakeand its aftershockzone are
earthquakeof 1769 [Richter, 1958]. Between containedwithin the area of activity in the
1963 and 1970, as reported by Caltech, three centerof Figures7a, b.
broad areas of activity in the San Fernando
1972 Bear Valley earthquake. The Bear
area are apparent (Figure 7b). Earthquakes Valley earthquakeof February 24, 1972 (M ----
8538
WEssoN AND ELLSWORTH; SEISMICITY PRECEDINGEARTHQUAKES
119Ol5 '
Fig. 7b.
5.0), occurredwithin a densenetwork of seismo- the aftershocksoccurring in the first 3 days
graphstations,affordingthe opportunityfor an
exceedingly
detailedlook at the distributionof
small earthquakes
precedingand followingthe
main shock[Ellsworthand Wesson,1972]. The
epicenterof the main shockand thoseof its
after the main event. The hypocenterof the
main shock, or the point at which rupture
began,occurredat a point on the fault surface
that had been the site of several preceding
events,includingtwo in excessof M -- 3.0 in
aftershocks
November
were
contained
within
one of the
most active segments
of this part of the San
Andreasfault zonefor the preceding3 years
(Figure 8). Earthquakesin the 17 months
preceding
the M -- 5.0 shockare shownin a
vertical sectionparallel to the San Andreas
fault in Figure9, alongwith the locationsof
1970.
The clusterof events11 km southeastof
pointA at a depthof 5-6 km represents
activity
associated
with a magnitude4.0 earthquake
December29, 1971.A plot of the numberof
earthquakeswith magnitudesgreaterthan or
equalto 1.0 per 2-by 2-km squareon the fault
WESSON AND ELLSWORTH' SEISMICITY PRECEDINGEARTHQUAKES
8539
121 ø 40'
:57o5
'
I
i
I/I/69
to 2/24/72
Magnitude
0/...
%
2.4
-
3.1
- 4.0
3.0
4.1 - 5.0
ß
Main
shock
Hollister
ß A .
M 5.0
ß
Zone
ofaftershocks
•c• •
ß
__
o
i
•
ß
City
ß
ß
ß
36o5 '
I
120
o 40
Fig. 8. Seismicity along the San Andreas fault system preceding the M -- 5.0. Bear Valley
earthquake after Lee et al. [1972a, b, c]. The locations of the main shock and the aftershock
zones are after Wesson et al. [1972].
surface [Figure 10] clearly indicatesthe occur- as in the Kern County earthquake,to the order
rence of the hypocenteradjacent to a concen- of a kilometer, as in the 1972 Bear Valley earthquake. The case is strongestfor this association
tration of precedingearthquakes.
in the situation
A CONCEPTUAL MODEL
where the data set is the most
complete and the locationsare the best (1972
In the seismichistories of the earthquakes Bear Valley earthquake).
These observations argue strongly that the
examinedin this studya cleartendencyemerges
for the hypocenterof a moderateearthquaketo stress distribution along typical fault zones in
occur within a relatively active region. This is California is sufficiently inhomogeneousthat, as
true on scalesranging from tens of kilometers, the generallevel of tectonicstressis increased
8540
WESSONAND ELLSWORTH' SEISMICITYPRECEDINGEARTHQUAKES
A
B
+
+
5
rrl
"o
-I
io
15
2o
0
----
I0
NW
A
20
:50
DISTANCE ( K M)
40
50
60
'•'
LONGITUDINAL
SECTION
ALONG
SANANDREAS
FAULT
9/I/70 TO2/21/72
SE
B
I
MAGNITUDE
•' i.o
_
1.0-
1.9
+
i 2.0
-2.9
+
3.0 - 3.9
4.0-4.9
2o
o
IO
LONGITUDINAL SECTION
20
30
DISTANCE
(KM)
40
ALONG SAN ANDREAS FAULT
15.56 2/24/72
50
TO 20:07 2/27/72
60
'•'
Fig. 9. Locationsof earthquakesprojectedonto vertical sectionparallel to the San Andreas
fault after Ellsworth and Wesson [1972]: (top) seismicity preceding M ---- 5.0 Bear Valley
earthquake and (bottom) first three days of aftershock activity.
(most probably as the result of plate motions),
relatively small areas are brought up to failure
beforea largesegmentof a fault fails.
A conceptualmodelof howthis process
might
occur,for example,alongthe San Andreasfault
near Bear Valley, is depictedin Figure 11. No
earthquakesare locatedat depths greater than
about 15 km, but the thicknessof lithospheric
plates is inferred to be much greater than 15
km; therefore it is reasonableto assume that
slip occurs without the generation of earth-
quakesbelowthis depth.Fault creepis observed
at the surface along some parts of the fault
[cf. Savageand Bur[o•d, 1973; Nason, 1971],
and it is also reasonableto assumethat, at
least locally, the zone of aseismicslip extends
to the surface. (This is not necessaryfor the
argument,becausesurface creep could also be
generated by the stressesdue to a pileup of
screwdislocationsbeneath a horizontally infinite
welded strip along the fault if the surface
strengthwerelow enough.)Alongpatchesof the
fault surfacethe physicalpropertiesor environmental factors (for example,coe•cient of friction or pore pressure)are such that aseismic
slip is at leastretarded(high coe•cientof friction or low pore pressure),perhaps as illustrated by the schematicdisplacementcontours
in Figure 11. Stresseswill be set up on this
patch as a resultof the internalstresses
of the
dislocation loops surrounding it, and the
strengthmay be decreased
by secondarynonlinear effectssuch as flow of pore fluids. Our
observationssuggestthat the differencebe-
WEssoN AND:ELLSWORTH'SEISMICITYPRECEDING
EARTHqUaKES
8541
A
B
IO
HYPOCENTER
i
•o
o
i0
1-5 •
i
I
30
DISTANCE (KM)
40
NUMBER OF
EARTHUAKES
='
6-10 !:•
I
I
20
50
i1
60
SE
11-15•
M
1.0
16ormore
Fig. 10. Number of earthquakesM •_ 1.0 per 2- by 2-km square plotted on vertical section
purullel to San Andreas fuult.
tween stress and strength are inhomogeneous the number of microfracturing events i.s obon the patch, leading to small earthquakes served just before failure. Our observations
where•he failurecriterionis met locally.Indi-
suggest
that,if thisanalogy
'holds,
theincrease
vidual small earthquakeswill lead to a spatial
smoothingof the differencebetween stressand
strength. This smoothingcan be accomplished
both by the 'priming model' of Weertman
in activity must occur over years to tens of
years. This sort of rapid increasein the number
of small earthquakes preceding a moderate
earthquake is not obvious before any of the
earthquakes studied, although tenuous arguments might be made for such increasein the
[1964, 1969], in which the internal stressesof
the dislocationsnucleated during slip are subtracted from the prestress,and by the flow of
pore fluids and resulting changesin strength,
as suggested
by Nut and Booker [1972]. Consequently, when the shear stresson the patch is
increased generally by continued or renewed
aseismicslip, a larger area will reach failure at
the sametime becausethe spatial irregularities
in the differencebetween stressand strength
are smoothed out. This results in a larger
earthquake. In this way the small e.arthquakes
that precedemoderateearthquakesare not only
symptomatic of circumstances favorable to
seismicslip; they may be a fundamental part
of the processleading to failure in a moderate
earthquake.
It is tempting to relate these observationsto
the results of small-scale experiments in the
laboratory that demonstratepremonitory phenomena prior to fracture or sudden sliding in
rock materials. Mogi [1962], Scholz [1968],
and others have demonstrated
that
microfrac-
turing occursin unbrokenrock prior to fracture
and prior to. frictional sliding on fracture surfaces. Generally, however, a rapid increasein
case of a few.
DISCUSSION
Moderate earthquakes in California seem to
occur in areas characterized by a relatively
high level of small-earthquakeactivity. Prior
to each of the earthquakes included in this
study, the region containingthe epicentercould
not be considered quiescent over a period of
years. Prior to most of the earthquakes, the
area immediately surroundingthe epicenterhad
a high level of activity relative to other segments of the same fault zone or other nearby
fault zones. These moderate earthquakes occurred in a variety of seismicenvironments.The
absolutelevel of small-earthquake
activity
varied immenselyfrom locality to locality, and
the moderate earthquakesincluded both thrust
and strike-slip mechanisms.The preceding
activity occurred predominantlyon the-same
fault as the moderate earthquake, but the
activity precedingthe BorregoMountain earth_quake appearedto occuron a parallel fault 10
km east of the strand along which ground
8542
WESSONANDELLSWORTH:SEISMICITYPRECEDING
EARTHQUAKES
[1959]. It seemsquite important that seismicity
studiesaimed at determiningthe level of small
earthquake activity be carried out in these
gaps.The analysisby Tobin and Sykes [1966]
of seismicityin Alaska for the 10-year period
prior to the Alaska earthquake of March 28,
1964, suggeststhat considerableactivity was
-15km
clustered
at each end of the aftershock
zone.
Fig. 11. Conceptual model of the vertical disIn particular, the activity in the northeastern
tribution of relative displacementu along a sec- end of the aftershock zone clustered near the
tion of a fault that slips by creep at depth and epicenter of the main shock. Kelleher and
arounda patch that slipsonly duringearthquakes,
Savino [1973] observed a concentration of
in arbitrary units.
activity in the vicinity of the epicenterpreceding severallarge earthquakesbut not preceding
breakageoccurred.In someof the areas,mod- the 1972 Sitka, Alaska, earthquake (M -- 7.6).
erate earthquakesare frequentevents; in other
Considerabledisagreementexists about the
areas they are infrequent. Moderate earth- relation between seismicity,strain, fault creep,
quakes have occurred near Parkfield several and geologicstructureand the potential for a
times in this century [McEvilly et al., 1967].
great earthquakein the apparent gap alongthe
In contrast, the San Fernando area had not
had a moderateearthquakein this centuryuntil
San Andreas fault in central California [Allen,
1971.
1968; Wallace, 1970; Scholzand Fitch, 1969,
1970, 1971; Savageand Bur•ord, 1970, 1971].
The occurrenceof a small earthquakeclearly This 175-km-long segmentof the fault, termirequires two physical conditions. First, the nated on the north by the fault breakageof the
stress in some region about the hypocenter 1906 San Franciscoearthquakeand on the south
mustexceed
the breaking
or frictional
strength by the fault breakage of the 1857 Fort Tejon
of the fault. Second,the nature of the breaking earthquake, is substantially more active than
or sliding surface, at least in the immediate either of the adjoining segmentsat the level of
vicinity of the hypocenter,must be conducive small and moderate earthquakes but has not
to the accumulation
and sudden release of
been the site of an historic great earthquake.
strain energy. These conditions are required Analysis of the strain and creep data [Savage
for the occurrence
of a moderateor largeearth- and Bur[ord, 1970, 1971] suggests
that strain is
quake as well. Some evidenceexists, however, not accumulatingat the present time within
that moderateand large earthquakes
may be
the resolution of the measurements.
complex,
multipleevents[Wyssand Brune,
The absence of strain a½cumalation at the
1967; Tri•unacand Brune, 1970],so that these
conditionsmay not be requiredover the entire
slip surface,but only on partsof it.
Fedotov[1965], Mogi [1968b,1969], Sykes
[1971], Kelleher [1972], and Kelleher et al.
[1973] have suggestedthat along primary
plateboundaries
suchasthe Kurile-Kamchatka,
present time (last few tens of years) may be
a relatively steady state phenomenon,as was
suggestedby Wallace [1970] and Allen [1968],
or it may be that the strain doesnot accumulate
linearly with time. For example, the rate of
strainaccumulation
mightdecrease
asthe effective stressapproaches
the ultimate strengthof
Aleutian-Alaska, and South American seismic the fault. The observationthat a high level of
zones,great earthquakes(M • 7.9) occur in small-earthquake
activity existsalongthis seg'gaps'betweenthe aftershock
zonesof previous ment does not distinguishbetween these two
great earthquakes.The earthquakesthat we possibilities.
It implies only that the difference
studiedincludedno great earthquakes,
but an
extrapolationof our observations
suggests
that
a greatearthquakewouldbe precededby a relatively high level of small and moderate earth-
between
the effective stress and the frictional
strengthapproaches•
zero.We cannotdistinguish
between a high level of shear stressor a low
frictional strength.To resolvethis question,we
quake activity. This has been suggestedfor require either (1) measurementsof stress and
the 1906 San Franciscoearthquakeby Tocher fluid pressure,(2) a better understandingof the
WESSON
ANDELLSWORTH:
SEISMICITYPRECEDING
EARTHQUAKES
8543
strength, or (3) evidence from the geologic
Many details of the relation between the
record on whether or not great earthquakes smallearthquakesthat precedemoderateearthhave occurredalongth!s segmentin Holocene quakesand the moderateearthquakesthemtime.
CONCLUSION
On the basis of these observations
of the
s9ismicityprecedingmoderateearthquakes
we
conclude
that
the
occurrence
of
substantial
small earthquake activity along a particular
fault zoneis diagnosticof (and in all probability
a fundamental part of) the tectonic processes
leadingto a moderate earthquakein California,
Althoughintuitive argumentscan be made suggestingthat a sustainedperiodo.fquiescence
is
required to accumulate sufficient elastic strain
selvesremain unsolved.However,with the increased density of seismographstations in
California,it is hopedthat thesequestions
can
be resolvedand that the potentialusesof this
informationfor the predictionof moderateand
larger earthquakescan be realized.
Acknowledgments.Wayne Thatcher, P•. M.
Hamilton, J. H. Healy, C. B. Raleigh, $. H.
Kelleher,and T. V. McEvilly read the manuscript
sandoffered numerous valuable suggestions.Dis-
cussions
with J. C. Savage,R. 0. Burford, R. A.
Page, and J.' P. Eaton were also stimulating.We
thank J. A. Hileman of the Seismological Laboratory, California Institute of Technology, for help
to generatea moderateearthquake,our obser- in obtaining a copy of the Caltech earthquake
vations suggest,and elementary physical argu- catalog on magnetic tape.
Research was carried out while one of the
ments support,the hypothesisthat local stress
concentrations
lead to the occurrence
of small
authors (R. L. Wesson) was pursuing an NRC-
earthquakesprior to moderateearthquakes.The USGS Post-Doctoral Research Associateship.
Parts of this work were done in cooperationwith
regional observations(with the possibleexcep- the Division of Reactor Development and Techtion of the Borrego Mountain earthquake) are nology, U.S. Atomic Energy Commission.
at least consistentwith, and in most instances
Publication authorized by the Director, U.S.
support, this hypothesis. The detailed obser- Geological Survey.
vations of the small earthquakesprecedingthe
]•EFERENCES
1972 Bear Valley earthquake are unequivocal.
The regional observations presented here Allen, C. R., Tectonic environment of seismically
active and inactive areas along the San Andreas
have potential implications for the analysis of
fault system, Proceedings of Conference on
seismicrisk. Moderate earthquakesin California
Geologic Problems of the San Andreas Fault
seem to occur along segmentsof fault zones
System, Stanford Univ. Publ. Geol. Sci., 11,
70-82, 19.68.
characterizedby a relatively high level of acAllen, C. R., Seismological environment, in the
tivity. This seismicityis high relative to
San Fernando, California, earthquake of Feb-
cent faults or segmentsof fault zones,which,
ruary 9, 1971, U.S. Geol. Surv. Pro[. Pap. 733,
7, 10, 1971.
on the basisof geologiccriteria for recencyof
displacement,
wouldbe considered
equallylikely Allen, C. R., and J. M. Nordquist, Borrego
Mountain eart.hquake: Foreshock, main shock,
to be the locationsof moderateearthquakes.
and larger aftershocks, U.S. Geol. Surv. Pro].
The problem arises, however, of a great
Pap. 787, 16-23, 1972.
earthquake,the hypocenterof which is along Allen, C. R., and S. W. Smith, Pre-earthquake
and post-earthquake surficial displacements,
an activesegmentof a fault, but whichruptures
into an adjacentquiescent
segment,
as suggested Bull. Seismol. Soc. Amer., 56, 966, 1966.
Allen, C. R., P. St. Armand, C. F. Richter, stud
by Weertman [1971]. Until the circumstances
J. M. Nordquist, Relationship between seisleading to great earthquakesare better undermicity and geologic structure in the southern
stood and capable of measurement,the usefulCalifornia region, Bull. Seismol. Soc. Amer.,
55, 753-797, 1965.
nessof an analysis of seismicrisk for moderate
Allen, C. R., A. Grantz, J. N. Brune, M. M.
earthquakes is limited.
Clark, R. V. Sharp, T. G. Theodore, E. W.
The detailed observations of the small earth-
Wolfe, and M. Wyss, The Borrego Mountain,
quakes preceding the moderate earthquakes California, earthquake of 9 April 1968: A prenear Bear Valley and the suggested
persistence liminary report, Bull. Seismol. Soc. Amer., 58,
1183-1186, 1968.
throughtime of smallareasof the fault plane
Allen, C. R., G. R. Engen, T. C. Hanks, J. M.
beingcloseto failure offer a potentialtool for
Nordquis•, and W. R. Thatcher, Main shock
mappingthis property alongfault zones.
and larger aftershocks of the San Fernando
8544
WESSONAND]•LLSWORTI-I'.SEISMICITYPRECEDING
EARTI-IQUAKES
earthquake of February 9 through March 1,
1971, U.S. Geol. St•rv. Pro•. Pap. 733, 17-20,
1971.
Bolt, B. A., and R. D. Miller, Seismicity of
northern and central California, 1965-1969,Bull.
Seismol.Soc. Amer., 61, 1831-1847,1971.
Bolt, B. A., C. Lomnitz, and T. V. McEvilly,
Seismological evidence on the tectonics of
central
and northern
California
and the Men-
docino escarpment, Bull. Seismol. Soc. Amer.,
58, 1725-1768, 1968.
Bonilla, M. G., Trench exposures across surface
fault ruptures a•ociated with the San Fernando
earthquake, in Geological and GeophysicalStudies, vol. 3, San Fernando.,California:,Earthquake
o• February 9, 1971, edited by L. M. Murphy,
National Oceanic and Atmospheric Administration, Washington, D.C., in press, 1973.
Brown, R. D., Jr., J. G. Vedder, R. E. Wallace,
E. F. Roth, R. F. Yerkes, R. O. Casfie, A. O.
Waanonen, R. W. Page, and J.P. Eaton, The
Parkfield-Cholame, California, earthquakes of
June.-August 1966--Surface geologic effects,
water-resources aspects, and preliminary seis-
mic data, U.S. Geol. Surv. Prof. Pap. 579, 66 pp.
1967.
Brune, J. N., a•d C. R. Allen, A micro-earthquake survey of the San Andreas fault system
in southern California, Bull. Seismol.Soc. Amer.,
57, 277-296, 1967.
Bulwalda, $. P., and P. St. Armand, Geological
effects of the Arvin-Tehachapi earthquake,
Earthquakes in Kern County, California, during 1952, Calq. Div. Mines Geol. Bull., 171,
41-56, 1955.
Byerly, P., Earthquakes off the coast of northern
California, Bull. Seismol. Soc. Amer., 27, 7396, 1937.
Clark, M. M., A. Grantz, and M. Rubin, Holocene
activity of the Coyote Creek fault as recorded
in the sediments of Lake Cahuilla, U•S. Geol.
St•rv. Pro•. Pap. 787, 112-130, 1972.
Dibblee, T. W., Jr., Geology of the southeastern
margin of the San Joaquin Valley, California,
Earthquakes in Kern County, California, dur-
ing 1952, Calif. Div. Mines Geol. Bull., 171,
23-34, 1955.
Dieterich, J. I.i., A deterministic near-field source
model, paper presented at Fifth World Conference on Earthquake Engineering, Int. Ass.
for Earthquake Eng., Rome, June 25-29, 1973.
Eaton, J.P., Instrumental seismicstudies, The
Parkfield-Cholame, California., Earthquakes of
June-August 1966--Surface Geologic Effects,
Water-Resources Aspects, and Preliminary
SeismicData, U.S. Geol. Surv. Pro•. Pap. 579,
57-65, 1967.
Eaton, J.P., W. H. K. Lee, and L. C. Pakiser,
Use of microearthquakesin the study of the
mechanicsof earthquake genera.
tion along the
San Andreas fault in central California, Tectonophysics,9, 259-282, 1970a.
Eaton, J.P.,
M. E. O'Neill, and J. N. Murdock,
Aftershocks of the 1966 P•rkfield-Cholame,
California, earthquake: A detailed study, Bull.
Seismol.Soc. Amer., 60, 1151-1197,1970b.
Ellsworth, W. L., and R. L. Wesson,Earthquakes
near Bear Valley, California (abstract), Eos
Trans. AGU, 53, 1042, 1972.
Fedotov, S. A., Regularities in the distribution of
strong earthquakes in Kamchatka, the Kurile
Islands, and northeastern Japan (in Russian),
Tr. Inst. Fiz. Zemli Aka.d. Nat•k SSSR, no. 36,
66, 1965.
Fedotov, S. A., On seismic cycle, po.ssibility of
quantitative seismic regionalization and longterm seismicprediction (in Russian), in Seismic
Regionalization of the U.S.S.R. edited by S. V.
Medvedev, pp. 121-150, Nauka, Moscow, 1968.
Hamilton, R. M., Aftershocks of the Borrego
Mountain earthquake from April 12 to June 12,
1968, U.S. Geol. Surv. Prof. Pap. 787, 31-54,
1972.
Kelleher, J. A., Rupture zones of large South
American earthquakes and some predictions, J.
Geophys. Res., 77, 2087-2103, 1972.
Kelleher, J. A., and J. M. Savino, Seismic preconditions of the 1972 Sitka earthquake and
several other large earthquakes (abstract), Eos
Trans. A GU, 54, 370, 1973.
Kelleher, J. A., L. R. Sykes, and J. Oliver,
Possible criteria for predicting earthquake locations and their application to major plate
boundaries of the Pacific and the Caribbean, J.
Geophys. Res. 78, 2547-2585, 1973.
Keylis-Borok, V. I., and L. N. Malinovskaya, On
regularity in the occurrence of strong earthquakes, J. Geophys. Res., 69, 3019-3024, 1964.
Lee, W. I.i.K., Earthquakes in the greater San
Francisco Bay region: 1969-1971 (abstract),
Eos Trans. AGU, 52, 815, 1971.
Lee, W. I-I. K., J. C. Roller, P. G. Bauer, and
J. D. Johnson, Catalog of earthquakes along
the San Andreas fault system in central California for the year 1969, open file report, 48
pp., U.S. Geological Survey, Washington, D.C.,
1972a.
Lee, W. I.i.K., J. C. Roller, K. L. Meagher, and
R. E. Bennett, Catalog of earthquakes along
the San Andreas fault system in central California for the year 1970, open file report, 73
pp., U.S. Geological Survey, Washington, D.C.,
1972b.
Lee, W. H. K., K. L. Meagher, R. E. Bennett,
and E. E. Matamoros, Catalog of earthquakes
along the San Andreas fault system in central
California for the year 1971, open file report,
67 pp., U.S. Geological Survey, Washington,
D.C., 1972c.
McEvilly, T. V., The earthquake sequence of
November 1964near Corralitos, California, Bull.
Seismol. Soc. Amer., 56, 755-773, 1966.
McEvilly, T. V., Magnitudes, epicenters, and
fault plane solutions,in the Santa Rosa earthquakesof O'ctober,1969, Miner. Inform. Serv.,
23, 44-48, 1970.
WESSON
ANDELLSWORTI•:
SEISMICITY
PRECEDING
EARTHQUAKES
McEvilly, T. V., W. I-I. Bakun, and K. B. Casady,
The Parkfield, California, earthquake of 1966,
Bull. Seismol.$oc. Amer., 57, 1221-1244,1967.
Mogi, K., Study of the elasticshockscausedby
the fracture of heterogeneousmaterials and its
relation to earthquake phenomena,Bull. Earth-
quake Res. Inst. Tokyo Univ., JO, 125, 1962.
Mogi, K., Development of aftershock areas of
great earthquakes,Bull. Earthquake Res. Inst.,
Tokyo Univ., •6, 175-203,1968a.
Mogi, K., Somefeaturesof recentseismicactivity
in and near Japan, 1, Bull. Earthquake Res.
Inst., Tokyo Univ., •6, 1225-1236,1968b.
Mogi, K., Somefeaturesof recentseismicactivity
in and near Japan, 2, Activity before and after
great earthquakes,Bull. EarthquakeRes. Inst.,
Tokyo Univ., 47, 395-417,1969.
Nason,R. D., Measurementsand theory of fault
creep slippagein central California, Roy. $oc.
N.Z. Bull., 9, 181-187,1971.
8545
in central California,J. Geophys.Res., 75,
4447-4453, 1970.
Scholz,C. H., and T. J. Fitch, Reply to comments
by J. C. SavageandR. O. Burford,J. Geophys.
Res., 76, 6480-6484,1971.
Sykes,L. R., Aftershockzonesof great earthquakes,seismicitygaps,and earthquakepredictionfor Alaskaand the Aleutians,J. Geophys.Res.,76, 8021-8041,1971.
Tobin, D. G., and L. R. Sykes,Relationship
of
hypocenters
of earthquakes
to the geologyof
Alaska,J. Geophys.Res., 71, 1659-1667,1966.
Tocher, D., Seismichistory of the..SanFrancisco
Bay region,SanFranciscoEarthquakesof March
1957,Calif. Div. Mines Spec.Rep. 57, 39-48,
1959.
Trifunac,M.D., andJ. N. Brune,Complexityof
energyreleaseduringthe Imperial Valley, California, earthquakeof 1940,Bull. Seismol.Soc.
Amer., 60, 137-160,1970.
Nur, A., a•d J. R. Booker, Aftershockscaused Udias, A., A study of the aftershocks and focal
by pore fluid flow?,Science,175,885-887,1972.
mechanism of the Salinas-Watsonville earthOakeshott,G. B., The Kern Countyearthquakes quakesof August 31 and September14, 1963,
in California'sgeologichistory, Earthquakes
Bull. Seismol.$oc. Amer., 55, 85-106, 1965.
in Kern County,California,during1952,Calq. Udias, A., Earthquake sequencesand seismicity
Div. Mines Geol.Bull., 171, 15-22,1955.
of the Watsonville region of central California,
Page, 1•. A., Crustal deformations on the Denall
Ann. Geofis.,22, 115-145,•1969.
fault, Alaska,1942-1970,
J. Geophys.Res.,77,
1528-1533, 1972.
Richter,C. F., Seismic
historyin the SanJoaquin
Valley,California,Earthquakes
in Kern County,
Unger, J. D., and J.P. Eaton, Aftershocks of the
October 1, 1969,Santa Rosa, California.,earth-
quakes (ab.stract), Geol. $oc. Amer. Abstr.
Programs, 2, 155, 1970.
California,during 1952,,Calif.Div. Mines Geol. Wallace, R. E., Earthquake recurrence intervals
Bull., 171, 143-146,1955a.
on the San Andreas fault, Geol. $oc. Amer.
Richter,C. F., Foreshocks
and aftershocks,
EarthBull., $1,2875-2890, 1970.
quakesin Kern County,California,during1952, Weertman, J., Continuum distribution of disCalif. Div. Mines Geol. Bull., 171, 177-198, locations on faults with finite friction, Bull.
1955b.
Seismol. Soc. Amer., 5•, 1035-1058, 1964.
Richter,C. F., Elementary
Seismology,
768pp., Weertman, J., Dislocation motion on an interW. H. Freeman,San Francisco,Calif., 1958.
face with friction that is dependent on sliding
Ryall, A., W. U. Savage,and D. B. Slemmons, velocity, J. Geophys. Res., 7•, 6617-6622, 1969.
Seismic potential in the western basin and Weertma•, J., Allen-Wallace zone of the San
range/easternSierra Nevada region, Nevada
Andreas fault and discontinuous yielding of a
andCalifornia(abstract),
Eos Trans.AGU, 53,
slip band in a crystal, Nature, 231, 9-10, 1971.
442, 1972.
Wentworth, C. M., and R. F. Yerkes, Geologic
Savage,J. C., and R. O. Burford, Strain acsetting and activity of faults in the. San
cumulation in California, Bull. Seismol. Soc.
Fernando area, California., The San Fernando,
Amer., 60, 1877-1896,1970.
Savage,J. C., and R. O. Burford,Discussion
of
paperby C. I-I. Scholzand T. J. Fitch, 'Strain
accumulation
alongthe San Andreasfault,' J.
Geophys.Res., 76, 6469-6479,1971.
Savage,J. C., and R. O. Burford,Geodeticde-
California, earthquake of Feb. 9, 1971, U.S•
Geol. Surv. Pro]. Pap. 733, 6-16, 1971.
We•on, P•. L., and D. H. Wilson, Faulting in
the San Fernando earthquake of February 9,
1971 (abstract), Eos Trans. AGU, 53, 444, 1972.
Wesson,R. L., P•. E. Bennett, and K. L. Meagher,
termination of relative plate motion in central
CataJogof earthquakes along the San Andreas
California,J. Geophys.
Res.,78, 832-845,1973.
fault system in central California, January-
Scholz,C. H., Microfracturingand inelasticde-
March, 1972, open file report, 60 pp., U.S.
Geological Survey, Washington, D.C., 1972.
Res., 73, 1417-1432,1968.
Wesson,R. L., J. C. Roller, and W. H. K. Lee,
Scholz,C. H., and T. J. Fitch, .StrainaccumulaTime-term analysis and geologic interpreta-
formationof rock in compression,
J. Geophys.
tion alongthe San Andreasfault, J. Geophys.
tion of seismic traveltime
Res., 74, 6649-6666,1969.
Rangesof central California,Bull. Seismol.$oc.
Scholz,C. H., andT. J. Fitch,Strainand creep
Amer., 63, 1447-1471,1973.
data from the Coast
8546
WESSONANDELLSWORTH:SEISMICITY PRECEDINGEARTHQUAKES
Wyss, M., and J. N. Brune, The Alaska earthquake of 28 March 1964' A complexmultiple
rupture, Bull. Seismol.$oc. Amer., 57, 1017-1023,
mated from the ratio of short- to long-period
amplitudes, Bull. Seismol. Soc. Amer., 61, 11531168, 1971.
1967.
Wyss, M., and J. N. Brune, Regional variations
of sourceproperties in southern California esti-
(Received May 24, 1973;
revisedAugust 10, 1973.)