ML20005A545

Testimony of Jn Brune Re Earthquake Hazard in Southern CA & Mexico,Strong Ground Motion in Mexico,Earthquake Mechanism & Strong Motion Along San Jacinto Fault.Prof Qualifications & Supportive Documents Encl.Related Correspondence

ML20005A545
Person / Time
Site:	San Onofre
Issue date:	06/22/1981
From:	Brune J CALIFORNIA, UNIV. OF, SAN DIEGO, CA
To:
References
ISSUANCES-OL, NUDOCS 8106300442
Download: ML20005A545 (128)

Text

._&} [ '

I 9 ,

, - = - . . . . _ _ - ,

l - .. _ . _ . _..

{ c) o I l c '.O

~l G ( g ,{ P '/N

~~

5 UNITED STATES OF AMERICA 9'j i I*

NUCLEAR REGULATORY COMMISSION * ".. '

BEFORE THE ATOMIC SAFETY AND LICENSING BOARE 0 IO3 I '* ~.~

\%

• My,"c2* .( ,

\(.sny Oc

. 's/

In the Matter of

/

I

",\ \

)

) Docket Nos. 50-361 OL SOUTHERN CALIFORNIA EDISON COMPANY, ) 50-362 OL ET AL. )

)

(San Onofre Nuclear Generating Station )

Units 2 and 3) ) 2 P O WRITTEN TESTIMONY OF <

'- Cs:hted JAMES N. BRUNE M g 1 -

@ JU:12 21981 > 3 I. INTRODUCTION J' y,

[y'[fy'p

.t::::n y ..

My name is James N. Brune. I am Professor $1 'e'o hY# s at the University of California at San Diego. My educational background includes a Bachelor of Science degree in Geological Engineering from the University of Nevada, and a Ph.D. in Seismology from Columbia University. In recent years I have carried out a number of studies relating to seismicity and tectonics of southern California and northwest Mexico, and to earthquake source mecha'nism and strong ground motion, particularly in southern California and Mexico. I am currently a principal investigator on contracts and grants funded by the United States Geological Survey and the National Science Foundation which pertain to earthquake hazard in southern California and northwest Mexico (USGS), strong ground motion in northws,*-

Mexico (NSF), earthquake mechanism and strong motion along the San Jacinto fault (USGS) and special studies of strong motion generated by the Imperial valley 1979 earthquake (NSF).

THIS DOCUMENT CONTAINS Sd g106300 M i [ POOR QUALITY PAGES 6# /

i l .-

a - -

i l

my biography and list of publications is attached hereto and incorporated herewith.

f e

0 0

l 4

j

o I ." a Ag i II.

SUMMARY

AND CONCLUSIONS l

(1) The state of our scientific knowledge is too limited to allow us to predict wi?.h confidence the maximum magnitude and consequent maximum ground accelerations to which the San Onofre Nuclear Power Plant may be exposed during its lifetime. The existing SER and FSAR reports do not counter these limitations. New and untested methods are introduced without adequate justification. Undefined and unscientific terms are used in conclusions. Assumptions are left unjustified and important data are not adequately considered. As a consequence of the above limitations, the adequacy of the design basis earthquake has not been established by scientific evidence.

(2) A prcposed design earthquake of magnitude Ms"7 is not the maximum which could reasonably occur on the OZD.

Based on the length of the OZD, our present understanding of the mechanics of earthquakes, the statistical distribution of earthquakes with nagnitude, and the tectonics and historic seismicity of the OZD_ Southern California-Northern Baja California region, it is quite possible that an earthquake of magnitude greater than 7.5 could occur along the OZD.

(3) The slip-raqe methodology introduced by the applicant is not valid for estimating the maximum magnitude earthquake on the OZD. It has not been verified by evidence, submitted to adequate critical review, nor established from principles of fault mechanics. The data cited in support of

D-C

• h\

l it can be explained by known statistical properties of occurrence of earthquakes, and the fact that only a short history of. seismic observations is available.

(4) Methods of estimating maximum magnitude based on some assumed fraction of rupture along a given fault zone have not been verified by evidence, submitted to adequate critical review, nor established from principles of fault mechanics. Choice of smaller fractions of rupture tength represent probabilistically more likely events and not maximum events in the sense of true mechanical limitations.

( .5 ) Because the data base of recordings of ground accelerations of large earthquakes (M N 7 and greater) at close distances (< 10 km) is very limited, and because our physical understanding of earthquake source mechanism and generation of strong ground motion is very limited, we cannot confidently establish expected values of ground motion for San Onofre from a major earthquake on the OZD. Different investigators have obtained significantly different results from correlations of peak ground acceleration with magnitude and distance, using essentially the same limited data base.

Present evidence indicates that peak horizontal and vertical accelerations could exceed the SONGS design earthquake values.

The probabilities of such exceedances are not known, but such exceedances have occurred several times in recent earthquakes.

(6) Numerical modelling procedures have not yet

_4_

r -_.

~

C 8 l

been de aloped and utilized sufficiently to be relied upon  :

in predicting ground motion, and have not at present been able to constrain expected values of ground acceleration significantly beyond the unreliable constraints arising from correlation of peak acceleration with magnitude and distance based on assumed curve shapes.

(7) The Imperial Valley earthquake of October, 1979 (IV-79), M, s 6.5, had a relatively low intensity compared to other earthquakes such as the Imperial Valley 1940 earthquake and the El Alamo 1956 earthquake. Nevertheless, it recorded peak accelerations and response spectra which exceeded the Design Base Earthquake for SONGS. Given the larger magnitude for the design earthquake at SONGS, the lower attenuation in the source region of the OZD, and the relatively low intensity and moment of IV-79, we may expect that exceedance of design values of acceleration and response spectra could occur at SONGS for an M = 7 earthquake on the OZD.

(8) The concept of " magnitude saturation" of observed values of peak ground accelerations with magnitude, previously hypothesized by Hanks and Johnson to apply between M = 5.5 and 6.5, is contradicted in the most recent studies of Hanks and McGuire, and the validity of its applicability for larger magnitudes is in doubt and not supported by data. Therefore, we may expect a general increase in average peak accelerations with magnitu'de above

, v--

o e M = 6.5, but the slope of the curve is not known at present.

i (9) The ground motion which might be caused by high stress drop earthquakes such as occur in Northern Baja California and in certain regions of Southern California have not been adequately considered in the FSAR and SER.

The 1956 El Alamo earthquake (M, = 6.8) generated a much greater area of intensity VI than the IV-79 earthquake, and this suggests that the near source ground accelerations were also considerably greater. Present understanding suggests that such high stress drop events, if they occurred on the OZD could produce ground accelerations considerably higher than the design ground acceleration.

E e

e 9

1

e-- -,

, o. * '

0, III. GENERAL STATUS OF ESTIMATING EARTHQUAKE HAZARD The state of our scientific knowledge concerning geology, tectonics, faulting mechanism, and generation of strong ground motion is too limited to allow us to predict with confidence the maximum magnitude and consequent maximum ground accelerations to which a critical facility such as the San Onofre Power Plant may be exposed during its life-

. time. The historic record of seismicity, our techniques of geologic investigation (especially in areas covered by water) and our understanding of earthquake mechanism, are all too limited to establish the earthquake potential of zcnes of deformation such as the Offshore Zone of Deformation (OZD) and the Cristianitos Zone of Deformation (CZD) 'which may pose a seismic hazard to the San Onofre Power Plant site.

Furthermore, once the magnitude for a design basis earthquake is decided upon, the data base for, and understanding of, the generation of strong ground motion, and in particular our knowledge of appropriate critical parameters such as fault dimensions, stress drop, rupture propagation, seismic wave attenuation, and seismic inhomogeneities, are too limited to predict with confidence the probable ground accelerations to be expected for large earthquakes (M near 7 and greater) s very near faulting (distance less than 10 km) .

Our lack of knowledge and the need for further research has been clearly indicated in the report of the Panel on

,y 3 m - m ,. _ ~ .y ,_m . . - . _ .- _

..y .. -._y.,-._,.. ,m., y _

o e Earthquake Problems Relating to the Siting of Critical.

Facilities of the National Academy of Sciences:

We do not now have the optimal information base that is required to. site all critical facilities to protect the citizens of the United States from the hazards posed by barthquakes--surface faulting, strong shaking, ground failure and tsunami. As a consequence, many facilities are

'overdesigned,' undoubtedly, others are 'underdesigned' to resist seismic effects.

Without the further research recommended by the panel, we cannot in general say which is which, and specifically in the case of San Onofre Nuclear Power Plant, whether or not it is underdesigned. As pointed out in their report:

Major gaps exist in our knowledge

• of seismic phenomena, and nowhere is this better illustrated than in attempts to'specify the locations, i frequencies and maximum sizes of future earthquakes that might affect critical facilities--the questions of 'where,' 'how often' and 'how big.' Seldom can all three of these questions be answered with anywhere near the confidence we desire.

In commenting on the specification of maximum earthquakes, the Panel comments:

Such events have been called the

' maximum credible earthquake,'

' maximum expectable earthquake,'

or, with regard to special facilities,

' safe shutdown earthquake' or simply the ' design earthquake.'

None of these terms has been precisely defined in a usable way, and what is ' credible' or

' expectable' to one person may not be to another.

l

e , -

Furthermore, the Panel says:

More often than not, in present practice, the concept of maximum credible or maximum expectable earthquake has some intrinsic elements of both scientific probability and acceptable risk.

In the opinion of the Panel, and I agree, this is not a satisfactory long term state of affairs, efforts must be made to separate the evaluation of scientific likelihood of a potentially disastrous event from the assignment of the risk that society is willing to accept for a particular critical facility.

These last sentences are particularly important in the present case because at critical pcints 'of conclusion both the NRC Staff and its consultants, and the Applicant and its consultants, use the term " conservative" or some similar term, without defining it, and also use the terms quoted earlier by the NAS Panel as terms not d precisely defined in a usable way."

This leads to uncertainty about the meaning of statements and in particular, uncertainty as to the level of risk assumed for society when a particular statement is being made.

Concerning the characterization of ground motions, the Panel comments:

...the statistical base of ground motion data is extremely limited....

At present, these estimates (of ground motion] are subject to considerable uncertainties, reflecting the limited historical data base and lack of detailed, quantitative knowledge of the influence of -

physical factors on ground motions.

Data are particularly limited for

near field and large magnitude earthquakes; unfortunately, such events pose the greatest hazards to structures.

The above quotes from the National Academy Panei represent a generally accepted evaluation of our lack of knowledge concerning earthquake hazard. Neither the Applicant and its consultants, nor the NRC 9taff and its consultants have presented evidence that these conclusions are inappropriate for considering the seismic hazard at San Onofre (SONGS),

and I believe that they must be taken as a base from which to judge any claims made by the Applicants or Staff to have scientifically established the conservatism and lack of risk associated with the Design Base Earthquake for SONGS. A general description of the evidence presented can be characterized as a compilation of evidence and arguments which support the theory that ground motion corresponding to the design base earthquake (DBE) are not likely to be exceeded, (at some unspecified level of prof 4ct$ity), while evidence and arguments to the contrary are a 2 aer~aately presented. There is no statement that the DEC cannoc be exceeded. The scientific evidence is mixed with non-scientific and undefined terms in such a way as to make the conclusions difficult to evaluate and of little use in arriving at a true sciantific understanding of the seismic hazard. New and untested methods are presented with inadequate scientific justification. Much evidence which seems to me to be almost irrelevant is presented in, great detail whereas critically important simple and true statements

~ , _ ' d such as "we do not know" or "we do not understand" are omitted, the result being that the truth is clouded and a misleading impression of undersEanding is given. It is clear that, with the burden of proof to show that the design basis earthquake ground motion could not be exceeded, the burden could not be scientifically held, given our present lack of knowledge.

e e

e

• e 9

.* rs hV. ESTIMATING THE MAXIMUM MAGNITUDE EARTHQUAKE ON THE OZD ~

. The NRC concluded in its Safety Evaluation Report (SER) :

The present evidence indicates an extensive, linear zone of deformation at least 240 kil'ometers long extending from the Santa Monica mountains to at least Baja California. We and our concultants consider this zone of deformation to be potentially active and capable of an earthquake whose magnitude could be commensurate with the length of the zone.

I assume that "an earthquake whose magnitude is commensurate with the length of the zone" means that the magnitude corresponds to an earthquake with rupture length corresponding to the length of the zone, since choice of any lesser rupture length would be less than the maximum length of rupture possible on the zone, and an arbitrary reduction in conservatism for which no basis is stated. Using the equaticn of Slemmons' (1977)

Table 13 relation for strike slip faults (Curve E) , a fault length of 240 km corresponds to a magnitude of 7.87. If we add one standard deviation, a magnitude of 8.6 is obtained.

If we assume a shorter length of 200 km for the OZD, as suggested by Slemmons (SER) , we obtain values of 7.76 (mean) and 8.45 (mean plus one standard deviation) . If we assume a magn'.tude corresponding to one half of the 200 km, we i obtain values of 7.35 and 8.05 respectively.

A 1967 report to Secretary of Interior Stewart Udall regarding the Bolsa Island Nuclear Power Plant statis, in the section entitled " Seismological Considerations" that

-1 2 -

~

In specifying the maximum earthquake for which public safety must be assured, a highly conservative approach has been adopted for two principle reasons:

(1) The consequences of some types of serious failure in a nuclear facility must be guarded against even if the likelihood is very remote; and (2) the historic record of earthquake occurrence is so.short that it cannot encompass the entire spectrum of possible events.

In view of the mandatory conservatism, we suggest that the maximum earthquake for which public safety must be assumed should be a magnitude 8 shock on the Newport-Inglewood Fault, or on one of the parallel offshore faults (Palos Verdes, San Pedro Faults).

Similarly, other studies have suggested an M = 7.5 and M = 7.25 as the OZD design magnitude (USGS Open Fild Report,81-115, (1980); Woodward-Clyde Consultants LNG Report, (1978)).

There is no physical reason why an earthquake rupture could not proceed along the whole length of the OZD.

No evidence is presented by the Applicant and its consultants or the NRC Staff and its consultants which suggests that rupture along the full length of the OZD is not possible.

However, this is clearly less likely than rupture along some fraction of the total zone, and thus a less conservative assumption would be that only a fraction of the fault would rupture, as proposed by the Applicant and its consultants and by the NRC Staff and its consultants. The choice of a smaller fraction (and consequently choice of a smaller magnitude) is a probabilistic choice with some greater. level.of risk implied.

p _ _

Slip Rate Method for Estimating Maximum Macnitude The use of the. geologic slip rate method proposed by WCC is not valid for estimating the maximum earthquake that could occur on the OZD, since there is no known reason why a fault zone of given length with a low slip rate cannot have as large an earthquake as a fault zone of the same length with a fast slip rate. Of course the probability of observing such an event in any given time period (e.g.,

the relatively short histor c time period over which magnitudes have been estimated) is lower the lower the slip rate. There-fore, use of such a method could be considered a procabilistic method for determining the maximum probable event in a given time period, and not a deterministic method for estimating the maximum magnitude which might occur at any time.

Figure 361.38-4 of the Woodward-Clyde Consultants *

• Response to NRC Questions shows a "Line Bounding Extremes

• of Bracketed Ranges of Data (MEL) . " This line is taken by the Applicant to represent the bounding curve which gives an M = 7.0 earthquake for the OZD, taken to have the same slip rate as NIZD, .5 mm/yr (no justification is given -

for assuming the bounding curve is a straight line rather than a curve). The sicpe of the bounding curve is controlled by i

only two points at slip rates below 1 mm/yr, and thus is quite uncertain. As indicated by Slemmons in the SER (p. E-7)

The data base for these figures is based on a very short historic record of earthquake activity; future earthquake and new data are I

14_

.-- +e

likely to extend the limits to some indeterminate .ti her value.

This can be appreciated by consider 6' ion of one point on the graph, namely, the point for the 1933 Long Beach earthquake.

If the magnitude for that event had been a little over one unit higher, M 3 = 7.5, the slope of the bounding curve would, of course, have indicated M > 7.5 as a maximum magnitude s

for a slip rate of 0.5 mm/yr and for the OZD. Thus, the slip rate method begs the question since it assumes a priori that the 1933 Long Beach earthquake is the controlling earthquake for a slip rate of .5 mm/yr. There is no justification for this because the historic record is too short. The 1956 El Alamo earthquake, discussed later, had a magnitude of 6.8, yet the slip rate average over the last several million years is less than for the NIZD. It can be noted that if we had had only two data points from the data above a slip rate of .5 mm/yr, we might have inferred a bounding curve with the opposite slope, i.e., maximum magnitude increasing with decreasing slip rate, a result which might be expected from a rock mechanics point of view, since it is observed in the laboratory that rock strength along faults increases with time between successive failures (Scholz, personal communication, 1981).

Anderson (personal communication,1981) and Luco (personal communication, 1981) have given simple interpretations of bounding curves, such as the MEL presented by the Applicant, in terms of a relationship between maximum magnitude, slip rate and recurrence time for the largest event. In such an s--~_ .

j l

l interpretation, the bounding curve could correspond to a recurrence time for the maximum earthquake of 2000 or 5000 years Thus, there are two possible interpretations of such a bounding curve: either there is some, at present unknown, mechanism operating such that no earthquake falls beyond the curve (as implied by the applicant), or more likely, the curve represents the result of limited sampling of seismicity for which the recurrence time for the maximum earthquake is of the order of 2000 to 5000 years, a reasonable situation for fault zones such as the OZD. -

1 Evidence from Historic Seismicity From the point of view of the historic seismicity of the OZD, an earthquake with magnitude greater than 7.5 is credible. Since an earthquake of magnitude 6.3 has airtsdy occurred on the northern section of the OZD (Long Beach, 1933) and a magnitude 6.8 earthquake has occurred on the San Miguel fault system to the southeast (El Alamo, 1956), and since it is very unlikely that the largest or near-largest event in the region would occur in such a short time sample as our historic record, given the relatively low slip rates, it is credible that a magnitude unit higher could occur at any time.

l

. l l

i

)

The San Miguel Earthcuakes and Maximum Magnitude Based on the Slip Rate Method Because of the proximity and the geologic and tectonic relationship of the San Miguel earthquakes to the OZD, they .

should have been given important weight in testing the slip rate method. The slip rate on the NIZD as given in Woodward-Clyde Consultants Appendix B is about 3.5 km in the last seven million years, giving an average slip rate of

.5 mm/yr (Figure B-7). On the other hand, the slip on the San Miguel fault zone in the last seven million years is only approximately 250 meters (Gastil, personal communication, 1981; Harvey, 1980) giving a slip rate of about .04 mm/yr.

According to the bounding curve given in the FSAR, Figure 361.38-4, this corresponds to a maximum magnitude of less, than-5.5, totally in disagreement with the magnitudes of actual earthquakes CM up t'o 6.8). This further supports the conclusion.that the bounding curve in 361.38-4 is a result of a sampling limitation, not a physical limitation on the magnitudes of earthquakes. If we take the value of .04 mm/yr (averaged over the last seven million years) for the 1956 earthquake, we obtain a bounding curve which gives a  ;

l magnitude of about 7.5 for the OZD (.5 mm/yr). l The calculated slip rate for the San Miguel fault depends on the time interval over which the averaging is done.

If the average is taken since Cretaceous, a value less than 0.01 mm/yr is obtained. On the other hand, if we assume o s the displacement all occurred in the last one million years, a rate as high as .25 mm/yr is obtained., The higher rates would be consistent with the thesis that the OZD-San .

Miguel linear zone is a highly active incipient fault zone.

Thus, the 1956 San Miguel earthquake is further evidence that the slip rate methodology is invalid for estimating the maximum magnitude for the OZD. The occurrence of an M = 6.8 earthquake in this region, on a fault of such low total displacement, and with such a short hiutoric record, argues that the maximum earthquake on the OZD could be considerably larger.

e l

l f

l

.s Fault Activity and Stress Away from the San Andreas Fault A number of times the suggestion has been made that solely because of its distance from the main plate boundary and from the San Andreas fault, the OZD is likely to have a lower stress or to have a lower slip rate, lower seismic activity and lower maximum magnitude (SCE Response to FOE Interrog. 39, 62, 64). Concerning stress, there is no mechanical basis for the assumption that stress decreases away from the plate boundary and, in fact, there are tectonic models in which the opposite is true (Lachenbruch, 1981).

The fact that the strength of faults increases with contact time in laboratory models suggests that faults of lower slip rate and thus longer intervals between earthquakes might have higher breaking strengths, higher stress drops and thus, higher magnitudes (for given fault geometry) than earthquakes along main fault boundaries. There are other suggestions that stresses away from plate boundaries are higher than those near plate boundaries (Sbar and Sykes, 1973) . The high stress drops and high apparent stresses found along the San Miguel fault zone (see later section) are strong evidence that in this region stress is higher away from the main plate boundary.

+

l

, , 1

' Magnitude by Fractional Fault Length 1

Since it is a well-known fact that for any given -

^

fault zone larger earthquakes are less frequent than smaller earthquakes , it is obvious thau the longer the observation time for a given fault zone the larger.the maximum recorded earthquake is likely to be and, consequently, the larger the 4 ra t, o of the largest observed ~ rupture length to the total fault-zone length,; i.e., the larger the fractional f ault length of the maximum event. For faults with slower slip rate, the observing time required to record a rupture with a given fractional fault length wil' be correspondingly longer than for a fault with faster slip rate.

l The definition of total length of zone is not straightforward, as pointed out by Slemmons (1977 and SER). For example, almost 100% of the Imperial fault broke in 1940.

However, we may extend the " fault zone" to which the Imperial 7

fault belongs to include, to tr_ north, the San Andreas fault, and, to the south, the whole set of transform faults in the Gulf of California, perhaps to the intersection wir.h the East Pacific rise, and thus make the " fractional fault length" for the 1940 rupture anywhere between necrly 100%

and less than 10%. Slemmons (SER) identifies the Imperial fault with the San Jacinto fault zone even though the San Jacinto fault is not believed to have been part of the main plate boundary in recent geologic time, while the Imperial e fault has. Furthermore,'the Imperial fault is offset from 40-

a simple linear projection from the San Jacinto fault to the Cerro Prieto fault. There are, at present, no objective rules for deciding before-hand what the proper length of the total fault zone is. Decisions about how to define the total length of a zone, and what fraction of the zone to take have considerable arbitrariness at present and the method has not been subjected to sufficiar critical review.

In the case of the OZD., the northern terminus is fairly clearly defined by the transverse ranges, but the southern end is less certain, as indicated by the various alternatives considered by Slemmons (SER). In fact, the

. right-stepping distance required to connect the Rose Canyon fault with the Coronado Banks-Agua Blanca system ($15 - 20 km)

. is about the same as that required to connect the Imperial fauit with the Cerro Prieto fault, two f aults

• which Slemmons (SER) includes in the same San Jacinto to Oerro Prieto zone. Thus, if we accept the Imperial and Cerro Prieto faults as part of the same zone for the purpose of estimating fractional fault lengths, there is no geometrical reason for not considering the Coronado Banks fault zone as part of the OZD.

For a given throughgoing linear fault zone, such as the OZD, there is no known reason why the rupture could not proceed along the whole length. Therefore, one might expect, given a long enough recording time, that each such more or less continuous straight zone would have increasing" observed f

l l

l l

l .

1

fractional ruptures with time, approaching eventually some

'**ge fraction of the total length. The 1930 IZU strike-slip

.rthquake ruptured nearly 100% of its length.(Allen, 1975).

Given the short time of historical observations and the inherent difficulties in defining total fault zone length, it appears possible that many relatively straight continuous active fault zones such as the OZD will eventually generate ruptures over nearly their whole length. In this context, the fractional fault length method appears as a probabilistic method, not a determinsitic one. Since a rupture of the whole fault length is less likely than rupture of some fraction of the length, assuming a rupture of length of some fraction of the fault length for a design earthquake is. simply less conservative than assuming rupture of the total fault length.

The probability of observing, in a given time period, a rupture along a given fraction of a fault zone can be calculated from the relationship between fault dimensions and magnitude and the recurreace relationship for :arthquakes of various magnitudes. Caputo (1973; personal communication, 1981) has made such a calculation and finds that the probability of observing one half length is about 4 times l

greater than full length, the probability of observing one third length about 8 - 10 times greater than full length, and the probability of observing one quarter length about 15 - 20 times grent'er than full length. From this consideration, it 1

is unlik<Ly, with the present data sample, that we.have ,

l recorded near the maximum rupture length possible on the OZD. I 5

i i

Estimating Magnitude from Rupture Length _

Slemmons (SER, App. E) has used a regression curve developed by Slemmons (1977)'to assign magnitudes to ruptures of a given length. In the calculations given by him in Appendix E, however, he uses the mean curve rather than the curve ts- a mean plus ' ne rcandard deviation. Thus, the magnitude values he cites for a given rupture length would be expected to be exceeded 50% of the time. The mean plus one standard deviation value is .694 magnitude units higher than the mean for strike-slip earthquakes. For example, for an assumed rupture length of 62 km (SER, p. E-ll) for SCOZD the mean estimated magnitude is 7.07 (expected to be exceeded 50% of the time), the mean plus one standard deviation is 7.77 (expected to be exceeded by about 16%

of the data for faults with a rupture length of 62 km) and i

the me'an plus two standard deviations is 8.46 (expected to 1

be exceeded by about 2% cf the data for faults with a j i

rupture length of 62 km). These calculations suggest that for rupture lengths of only a fraction (%1/4 to 1/3) of the j length of the OZD an M = 7.5 event is possible.

1

~

I l

l Synthetic Slip Rate v. Magnitude Cur're Section 361.38 (b) in the FSAR is a comparison of the slip rate and half-length methods for estimating maximum magnitudes.. This is done using a synthetic slip rate vs.

magnitude plot based on two correlations: the magnitude vs. rupture length correlation of Slemmons (1977) and a correlation of slip rate vs. length (Fig. 361.38-3) to obtain a synthetic one-half ' length line (Figs. 361.38-4. '61.38-5, 361-38.6). However, both of these correlations represent average values, and thus the synthetic slip rata vs., magnitude

, plot also represents an average line. If the data of Slemmons (1977) for strike slip faults is transferred in the same .

manner, 50% of the data will fall to the right of the curve,

• indicating that the bounding curve from the slip rate does not " bound" the data. A more conservative estimate would include a one st'andard deviation correction (t.694 magnitude units) giving a maximum magnitude of about 7.35. A two standard deviation correction would give a magnitude value slightly over 8.

Strictly speaking, in order to derive a valid " bounding curve", a bounding line should have been drawn for the data in Fig. 361.38-2, combined with a bounding curve for the magnitude vs. rupture length data, and transferred to Fig.

361.38-3. This curve would indicate a magnitude of about 8.5 for a slip rate of .5 mm/yr.

p--w- -

.. .1 Japanese Earthquakes -

In'FSAR Question 361.46b it is stated that "The differences in mechanics of faulting between Southern California and Japan has led the Applicants to remove the Tanna fault from the data base together with eliminating all Japanese faults from consideration."

Removing the Japanese-data from consideration has a serious effect on the conclusions concerning the slip rate method. Since the Japanese data represent most of the data at slow slip rates, the data base is weakened in precisely the range where it is most uncertain and where the data is most important to the conclusions concerning the maximum earthquake limit (MEL). Since much of the Japanese data exceeds the present proposed MEL, its climination has shifted the MEL curve to lower magnitude values for slow s_1p rates.

Considering the claims made for the slip rate method by the Applicants and the NRC Staff, it is important to thoroughly justify such dismissal of data. There is no established reason why Japanese strike slip earthquake mechanics should be any different than California strike slip earthquake mechanics.

Differences between Japanese earthquakes and Southern California earthquakes mentioned in the FSAR jnclude:

Most strike-slip faults on land in Japan began in the early i

~

Quaternary (about one million years ago) and have continued to move in the same direction with an average rate of a few millimeters per year (Matsuda and Okada,-1968). The total displacements are not greater than 12 kilometers. This small amount of displacement and youthfulness of the origin of their recent movement are characteristic of Japanese active faults and are in contrast with the history of such major faults as the San Andreas fault in California or the Alpine Fault in New Zealand (Matsuda, 1967). FSAR Question 361.46 b-2 Recurrence intervals of earthquakes on a given fault are long, and lie in-the range of several hundred to a thousand years (Matsuda, 1967). FSAR Question 361.46 b-2 Japanese strike slip earthquakes differ from all other strike slip earthquakes in many ways. The most conspicuous differences are the large displacement relative to rupture length, the shortness

. of rupture. length and overall fault length, the fact that commonly the entire mapped length of the fault breaks in one earthquake and that conjugate pairs of faults often rupture at the same time. FSAR Question 361.46 b,2 There is a trend of apparently low slip rates for some strike slip faults in Japan that produce large earthquakes.

FSAR Question 361.50.

To the extent that strike slip faults in Japan and California fall into almost mutually exclusive groups when i these fault properties are compared, the Applicants conclude that faulting of the kind that occurs in Japan cannot occur in California. Accordingly, it is inappropriate to include faults in Japan in an analyses of fault behavior of strike slip faults in California.

FSAR Question 361.50. .

Apparently, the Japanese data are disregarded because they commonly violate the thesis that is being tested. The data set for Japan is probably the best in the world, in terms of thoroughness and length of historic record. This, in part, may be the reason for its appearing anomalous to the Applicants. It seems to me premature to disregard the Japanese data until some mechanism is established justifying this, or until better data are obtained cutside Japan. It may be noted, that many of the characteristics of Japanese earthquakes which are used to justify disregarding them are~

precisely those characteristics attributed by the Applicant to the OZD, e.g., youthful origin, low slip rates, small total displacements and long recurrence intervals on a given fault. Thus the Japanese data should not be disregarded in considering the seismic hazard f* rom *the OZD. Similar comments apply to data from Chi'nese earthquakes, which are dismissed. There are examples of Chinese earthquakes of large magnitude occurring on fault zones which have remained quiescent for periods much longer than the period of observation available for the OZD (Allen, 1975).

l l .

l l

l 1 i

V. PEAK GROUND ACCELERATIONS The Data Base for Estimating Peak Ground Accelerations The data base for predicting peak accelerations very near (< 10 km) from large earthquakes (ftu7 and greater) is very limited, and thus any such predictions are open to large uncertainties. We will need about ten well-recorded earthquakes of magnitude near 7 and greater before we can be confident of such predictions. Present data do not preclude occasional peak horizontal accelerations of higher than 1 g at a distance of 8 km.

A similar conclusion (regarding possible peak horizontal acceleration) results from consideration of our physical understanding of earthquakes and generation of strong ground mo tion. - We lack the necessary understanding of critical aspects of the rupture mechanism, e.g. level and variation of stress drop, complexity of rupture, focussing of energy by rupture propa'gation, and attenuation. Simple theories relating to peak acceleration to stress drop indicate accelerations higher than lg are possible for stress drops of 100 bars, a value reasonable for eartnquakes of magnitude 7.0. However, localized stress drops of higher than 500 bars have been inferred in some studies, thus, since peak acceleration is linearly related to stress drop in these simple models, near cource accelerations of higher than Sg might occur. It is not known at the present time,,

i because of the lack of data, how adequate our present simple 1

l l

models are for estimating ground accelerations, and it is not known what the probabilities of high stress concentra-tions are, nor how large a volume can experience large stress drops, and, thus, how far very high accelerations could extend away from the source of the concentrated stress drop.

Because of the uncertainties described above, it is not possible to establish with confidence the probabilities of high peak horizontal accelerations at a distance of 8 km from an M = 7 earthquake. Recent extrapolations of existing data obtain quite different results (TERA Report Campbell, 1980; NRC SER; USGS81-365, Joyner, at al., 1981). The differences in the estimates stem from small differences in choices of data base and in assumptions in the regression analyses. At present, it is not possible to establish with confidence which data and which assumptions are most appro-priate and thus, the probabilities of high accelerations remains uncertain. From the results of the USGS Open File Report 81-365 correlation, the probability, for an M = 7 earthquake at 8 km, of accelerations over lg is 11%, and for accelerations over 2g, 1%. These values are higher than for the other correlations but do not appear to be unreasonable, based on our present undarstanding of the data and of earthquake source mechanirm. However, they must remain quite uncertain because of the limited data base.

. 1 s

t Possibility of high stress dron For other fault parameters constant, accelerations and velocities are proportional to stress drop. The average stress drop for large earthquakes is about 30 bars (Kanamori and Anderson, 1975) with a range up to about 100 bars.

Although most small earthquakes have stress drdps of less than 100 bars, there is evidence from spectrum studies that in some circumstances stress drops can be as high as a kilobar, with consequent higher nearly source accelerations and velocities. Trifunac (1972 a and b) found stress drops of about 500 bars and 350 bars for aftershocks of the San Fernando and El Centro earthquakes, respectively. A stress drop of N600 bars was inferred by Hartzell and Brune (1977) for one earthquake in Brawley earthquske swarm of January 14 - 31, 1975.

Fletcher et al., (1978) fo'und a stress drop of over 400 bars for the Oroville, California ear:hquake of August 6, 1975.

Hartzell et al., (1978) inferred a stress drop of over 1 kbar for the Acapulco, Mexico earthquake of October 6, 1974. The larger earthquakes in a recent Victoria, Baja California swarm apparently had stress drops of over 500 bars. Other examples of studies which found high stress drops include House and Boatwright (1980), 890 bars and 650 bars; and McGarr (1981 over 2 kbar for asperity stress drop).

The stress drop along a major fault during a large earthquake is probably quite variable, and thus even though the average stress drop is usually less than 100 bars, locally the stress drop could be considerably higher. Hanks (1974) inferred a stress drop of 350 - 1400 bars near the hypocenter of the 1978 San Fernando earthquake. Aki (1978) inferred a w

local stress drop of 370 bars and associated near source i accelerations of 1.5g for the 1857 California earthquake (based on variations of observed fault slip and a barrier theory of faulting). Trifunac (1972a) inferred a stress drop of about 350 bars for the southeast part of the rupture in the 1940 Imperial Valley earthquake. It is not known how large the areas of high stress drop along a fault can be, but it is possible that they could extend over more than a 10 km radius, thur leading to anomalously large accelerations and velocities at the surface and out to 10 km from the fault. Although our understanding of earthquake stress drop is not fully developad, and there are uncertainties in the above results, th*y e collectively suggest that in some cases stress drops of a few hundred bars or more may occur over fault volumes of at least a few km dimension'.

Whether or not large stress drops can extend to shallow depths (less than a few km) is not known. The study of Aki (1978) implied large stress drops extending to.near the surface. Hartzell, et al. (1978) interpreted the strong motion record of the Acapulco earthquake of October 6, 1974 to be due to an earthquake with a stress drop of over 1 kbar at a depth of only about 1 km. The Norma 163 model of l

Heaton and Helmberger ( 1979) for the San Fernando earthquake ,

I l

l y _

__7-__. y

has two fault segments one at a depth of about 13 km with a relatively steep dip and a second at depths less than 5 km with a shallower dip. His contours of displace-ment on the shallower segment of the fault imply. high stress drops at 2-3 km depth. Thus, there are indications that in some cases stress drops greater than a few hundred bars may extend to near the surface.

I conclude that large stress drops over relatively large volumes near the surface could cause anomalously high accelerations and velocities in some instances (greater than 2g accelerations and greater.than 200 cm/sec velocities).

The probabilities of occurrence for high stress drops is not known.

l Effect of focussing of energy by rupture propagation (Directivity)

Focussing of energy in the direction of source propa-gation is a phenomenon that has been known and observed in nature for many years. In seismology, the effect has been termed directivity and has been observed for many earthquakes (Bakun, et al. 1978), and most recently in the Livermore earthquake (Boore e t al. , 1980), the Santa Barbara earthquake (CDMG, Special Report 144, 1979; Miller and Felszeghy (1978);

and the Coyote Lake earthquake ( Archuleta, 1979).

32- #

e, .

nt

1 For wavelengths shorter than fault dimensions, the effect can lead to amplitudes in the direction of rupture several times higher than in directions away from (or near normal to) the direction of the rupture. The effect has been verified in physical models of spontaneous rupture (Archuleta and Brune, 1975) and in numerical similations (Boore, 1977; Archuleta and Frazier, 1979) as well as in numerical modelling of TERA-DELTA (page 6-1 , Supp. 1 ).

The concept of focussing or directivity is bmportant in strong motion seismology not only because of the fact that it can lead to anomalously high ground velocity and acceleration in the focussed direction, but also because it can introduce a large range of scatter in the data close to faults, thus ma. king it particularly difficult to estimate -

the true mean and standard deviation of peak velocities and accelerations from a limited sample of data. It may cause the distribution of " ax accelerations to deviate signifi-cantly from a legnormal distribution.

Although, as mentioned above, effects of directivity have been observed for several recent earthquakes, there is no case of a well instrumented large earthquake (M g near 7) where such effects are clearly evident (directivity effects are not obvious in acceleration data from the recent IV-79 earthquake, possibly because the source was not an approximate uniform rupture). Thus, the possibility remains that if

- 7 r

special circumstances leading to strong directivity for a large earthquake were_to occur, horizontal accelerations could be considerably higher than any recorded to date.

The probabilities of such occurrences are not known.

Arguments against high velocities and accelerations I have discussed above, a number of points which suggest that near large earthquakes accelerations higher than lg and velocities higher than 100 cm/sec may be common, and accelerations as high as 2g and velocities as high as

.200 cm/sec are possible. I would now like to discuss some of the arguments which have been cited against the possibility of such high' velocities and accelerations.

The fact that the data base is so small can be equally well used to argue tha't the abovh conclusion, that high velocities and accelerations will be occasionally expected, is not proven (burden of proof reversed), especially since no accelerations as high as 2g nor velocities as high as 200 cm/see have yet been recorded. Also, a number of physical phenomenon might limit the velocities and accelerations observed, e.g. scattering, inhomogeneities in the rocks, incoherency in the fault rupture, low Q and high non-linear  ;

l attenuation. Non-linear attenuation at high strains assocla- .

l ted with high acceleration might be especially effective in I i

j g

limiting accelerations in certain types of soil such as exists in the Imperial Valley. The fact that average stress drops (averaged over the fault plane) are commonly about 30 bars, and thus less than necessary for generating large accelerations and velocities, suggests that in most cases such large velocities and accelerations would not be expected (probabilistic argument) . Also, perhaps many of the recently observed high values of acceleration and velocities have been unduly affected by special conditions which would not apply to San onofre.

Building damage observed near large earthquakes has usually not been ,as great as engineers would have expected-for such large accelerations and,velocitics. This observa-tion suggests that such large accelerations and velocities are rare, if we accept these earlier expected correlations of building damage and acceleration. However, estimation of ground acceleration from building damage is a very un-certain procedure. Conversely, however, this line of reasoning, along with the low intensities associated with the accelerations recorded for IV-79, suggest that higher intensities, e.g., intensity IX, may be associated with higher accelerations than previously thought.

Finally, it can be reasonably argued that the very high values of accelerations and velocities require such a coincidence of deviations of variables away from their l

- 3 5-l

< *

• j average values as to be very unlikely for any given earth-quake (probabalistic argument) . For example, the highest accelerations are expected when a series of factors combine to lead to high accelerations (e.g. station azimuth and rupture propagation so as to maximize directivity focussing, anomalously high stress drop, relatively low attenuation and scattering, and anomalous surface amplification).

In my opinion, the above arguments do not outweigh the contrary arguments and evidence. They are especially weak if the burden of proof is assumed to lie with the contention that high velocities and accelerations are not possible.

Expectations for the near future

, It is evident that our understanding of the nature of strong ground motion near large parthquakes is still in an uncertain stage. Deployment of large numbers of accelero-graphs near active faults began only a few years ago and '

the data base is as yet very limited. Each new large earth-quake recorded usually has surprises. We may expect marked changes in our ideas once strong motion from several large earthquakes has been observed on a number of instruments in the near field. Also, our ability to do theoretical and numerical modelling is advancing rapidly and may lead to important insights in the near future.

Conclusion regarding the data base for gratnd motion Based on our present limited data base for near source

(< 10 km) ground motion for large earthquakes (M S 7), and based on our present limited understanding of the seismic l

l wave generation and transmission, the design base horizontal ground acceleration of .67g could be exceeded by an M = 7 earthquake on the OZD. Under, reasonable assumptions, maximum accelerations at a distance of 8 km could exceed ig.

Although there.are factors operating which might make large accelerations and velocities less probable, such limiting factors are not established by our present data base and theoretical understanding. A near certain conclusion is that if the burden of proof is assumed to lie with the thesis that very close (< 10 km) to large earthquakes (M N 7 and greater) accelerations of greater than .67g are not common, then the thesis has not been proven.

4 6

e l

l l

__I

1 Numerical Modelling Numerical modelling provides a method of extrapolating beyond our present data base to magnitudes and distances not represented in the data. This can be useful in und'erstanding the possibilities and probabilities of high ground accelera-tion. However, to do this would require variation of the input parameters over ranges based on outside information.

Since by itself the ground acceleration data at close distgaces and large magnitudes is so limited that different investigators can come up with quite different' extrapolations beyond the d:ta base, as described above, it is clear that, given the I flexibility of assigning input parameters in the numerical modelling, it is also possible for different investigators -

4 using numerical modelling to obtain different conclusions -

about whether or not the 0.67g DBE for San Onofre is

" conservative". 'The modelling studies of TERA-DELTA purport to show that the .67g DBE horizontal ground acceleration for San Onofre is " conservative" and to provide support for this conclusion beyond that provided by " empirical" t

study of the data base. I disagree with that conclusion because the study does not demonstrate that the parameters

, introduced into the numerical modelling are " conservative" l and have been. varied over reasonable ranges.

Rather, the l study shows that the parameters can be cho;Sn in such, a i

I j,

p

4 l

l i'

way that the resultant accelerations and response spectra fall below the values for the DBE (at the appropriate distance and magnitude) and at the same time fit reasonably well limited data from other distances and for ott.er magnitudes. It is not surprising that such a " solution" (i.e. selection of parameters) can be found, given the number of modelling parameters introduced, their uncer-tainties, and the lack of controlling data in the distance and magnitude range of interest. Unfortunately, such a result does not help very much since we already knew that a set of assumptions could be made, in extrapolating from the existing limited data base such that the DBE acceleration of

.67g would be above the extrapolations.

However, the real value of nurterical modelling is not taken advantage of in this approach. I believe that the proper approach is to assume a range of various input parameters, the ' range determined by information other than the strong motion data itself. Thus, parameters which are introduced for which there is little knowledge about their true values, would result in large variations in the result, but effort would be put into Ibniting their range using outside information. Given our present uncertainties in many of the parameters, it is obvious that many of the computer runs would.show results over the .67g DBE value.

This, by itself, would not necessarily mean that tha DBE

acceleration was not adequate or conservative. However, we would be in a position to judge better the likelihood of these exceedences. Unfortunately, such calculations are so expensive that few organi=ations have the funds to carry them out. Given the fact that such a variation in parameters has not been carried out, I have to agree with Drs. Trifunac and Luco, in t:teir testimony before the MRC Appeal Board hearings on the Diablo Canyon Nuclear Power Plant (October, 1980) that use of such calculations in the licensing process is premature. In the present case, with the present stage of development, the calcula-tions presented by TERA-DELTA do not add a significant further constraint on expected values for ground motion over that available from statistical extrapolations from the data base itself

, With these introductory remarks about the philosophy of numerical mod'elling, I sould like to address the uncer-tainties of some of the parameters in the TERA-DELTA study.

First, the values for standard deviations in the TERA-DELTA model do not represent the kind of standard deviations expected from real data where stress drops, rupture complexity and rupture propagation, as well as relationship of the rupture to local geologic irregularities, varies from earth-quake to earthquake. In the r(al world, standard deviations of peak acceleration are considerably greater than the values 5

given in the TERA-DELTA study, which correspond to varying only certain randomness parameters in a controlled way.

Second, the effects of uncertain values of the atten-uation parameter Q have not been adequately investigated.

NRC Reviewer Luco (by report to the Nuclear Regulatory Commission, August, 1980) suggested they should be about a factor of 2 higher. I believe this is a reasonable factor for the uncertainty of Q. Low Q values assumed in the TERA-DELTA model may have excessively attenuated high frequency energy and thus reduced peak accelerations (as well as indirectly and artifically reduced the effects of focussing in the modelling).

Third, because of the assumed slip function in the TERA-DELTA model, it is difficult to infer what the dffective value for dynamic stress drop is. A dynamic source study comparison by Swanger et al., ti981), indicates l

that it is only about 50 bars (see attached fiture). (There is no established basis for the assumption, made in the TERA-DELTA modeling, that the peak slip velocity parameter is constant for all earthquakes.) I feel that the average effective stress values should be varied by at least a factor of two and, in addition, the possibility of localized stress drops of up to 500 bars should be considered. The parameter studies by TERA-DELTA have shown that in the frequency range 2 to 10 Hz, the response spectra are essentially line,ar with with respect to dynamic stress drop. Thus, to take into 100 -

G7NAMIC MODEL

~

_ TERA / DELTA 100EL (V, = 8, TR " Z'9) g100 _

e -

t )' N -

w

• %,'j//+f M E

~

/ q .h ,

~

p h

, sf ,'

w 9

y a /

j a

g e* -

5 s  :

5 . @#

c .

1 . ......I , ,e e i .e I , , .

0.01 ,

0.1 1.0 PERIOD (sec)

Figure 4.7. Comparison of response spectrum of the dynamic slip function with that of the TERA / DELTA San Onofre slip function.

SWANGER, ET AL.

account these higher stress drops, we should multiply the spectra in this frequency range by a factor of about two.

Local zones of high stress drop would increase the spectra even more.

Fourth, the TERA-DELTA model does not adequately .

predict the accelerations actually observed in the Imperial Valley 1979 eart!. quake at stations a few kilometers from the fault, being approximately a factor of two too low (Luco, 1980). This indicates that the dynamic stress drop or some other parameter should be changed to increase the peak acceleration by atout a factor of two in order to fit the Imperial Valley earthquake. Further increases would 4

need to be made to go from the IV structure to the SONGS structure and to extend the results to the higher magnitude appropriate for the OZD.

In conclusion, the results of the computer modelling presented by TERA-DELTA indicate the feasibility and importance of doing a parameter study of expected strong motion at SONGS but, by themselves, do not constitute such a study, but rather, a limited range of calculations in whir.h important parameters have been assigned values not established as conservative. If these parameters had been assigned more conservativo values (based on our present uncertainties) the DBE spectra would have been substantially exceeded.

Requirements for an adequate modelling study An important aspect of. estimating earthquake hazard with the present limited data base is the use of computer modelling of fault rupture to simulate strong motion, with-out which we cannot make full use of the data. Modelling can provide improved understanding of the strong motion, especially the effects of focussing by rupture propagation and localized concentrations of high stress drop. Modelling can eventually significantly reduce uncertainties and peovide a range of realistic models of ground motion records from which to predict expected ground motion which might occur at SONGS from a rupture on the OZD. To estimate the effects of uncertainties in the parameters for the SONGS modelling,

,, reasonable variations in model parameters should be carried out to indicate conclusions for a range of degrees of conservatism.

The numerical modelling should include effects of rupture on all faults considered capable. In particular, the orientations and locations of fault ruptures should be varied to take into account possible effects of focus sing, i.e. whether fault rupture could proceed toward the plant, thus focussing energy in that direction. A conservative approach in this case would be to choose fault orientations as close to the direction toward the plant as allowed by the data.

A distance of about 5 miles or 8 km with an uncertainty of the order of one kilometer has been cited by the Applicant and Staff as the distance to be considered for the design earthquake. However, at the time this decision was made, detailed seismic sounding information did not exist between

-the OZD and SONGS. A recent study by Greene and Kennedy (1980) shows a zone of deformation, called the Cristianitos Zone of Deformation (CZD) lying closer to SONGS at a distance of about 1500' offshore (2 miles or 4 km). Since northwest of the zone is a " data void", it is reasonable that the zone may continue parallel to the coast and simply be a closer strand of the OZD (rather than a branch of the Cristianitos fault from onshore). Both Greene and Kennedy (personal communication 1981) have indicated that this is a reasonable interpretation of the existing data. If'this possibility is established as credible, modelling should include calculations with rupture on this znne. It is also important to consider the effect of possible rupture on any other splay or branch faults considered capable. In particular, calculaton should be made for a rupture proceeding northwestward along the OZD and continuing onto the CZD towards SONGS. Focussing and directivity effects of such a rupture might lead to peak accelerations at SONGS considerably greater than the DBE.

A reasonable initial conservative model for an earth-quake on the OZD would be a model with a more or less uniform stress drop of 100 bars (over the entire fault rupture), and I

l

superimposed local stress drops of about 500 bars for local

, stresa concentrations of about 5 km in radius, located at several points along the main fault" branch and on splay or branch' faults which are judged capable. Refined estimates for these values could be made when the initi.al results of the modelling are obtained.

A possible criticism of this proposed variation in input parameters for numer. cal mcdelling is that we already know from existing sensitivity studies and the large uncer- .

tainties (due to lack of controlling data) in input para-meters, that such calculations will in some cases yield accelerations higher than the DBE, and thus, we do not need to spend the money. Nevertheless, it is clear that more sensitivity studies (in essence the same as variatiIon in parameter studies) are needed before the full value of numerical modelling is realized.

1 Lessons from the Imperial Valley 1979 Earthquake The Imperial Valley 1979 (IV-79) earthquake is the

'first well-recorded (in terms of near source strong motion) earthquake with M, between 6 and 7. Accelerograms were recorded at sites along the fault trace, at sites in both directions from the epicenter, and on a cross array of stations near the northern end of the fault. Because it is unique in terms of instrumental coverage, there is a natural tendency to assume it is the " typical" earthquake for its magnitude, and use it as a basis for estimating strong ground motion for other magnitudes and other geologic settings.

However, because the data base is so limited, we cannot confidently assume the IV-79 earthquake is " typical" or "not typical." We will.need several more well-recorded earthquakes before we establish with certianty how " typical" it is. Furthermore, it will probably be at least a couple of years before the mechanism of the IV-79 earthquake is analyzed and understood well enough to clearly appreciato its implications for strong motion.

Horizontal accelerations near or above the San Onofre DBE value of .67 g were recorded at several stations: 72 g at Station 942, .81 g at Station 5054, .64 g at Station 958 and .61 g at Station 955 (Data taken from Table 2 of Joyner

.et al., 1981). Thus, the IV-79 earthquake data suggest that near source values of horizontal acceleration above .67 g may be quite common for earthquakes of this magnitude. There is no reason to expect that such high values could not occur at San Onofre from a similar earthquake on the OZD. In fact, the TERA-DELTA modelling results indicate that for the same earthquake mechanism, the accelerations at San onofre would be about 1.8 times higher, primarily because of lower attenuation in the SONGS structure as compared to the IV structure (Final Report, Fig. 4-13, attached).

t 5

l 1

l

. l

~

C GREEN'S s FUNCTIONS 5 t

Source Depth 5 krn E Distance 21 km [ j o

t e

3 5

VERTICAL AZIMUTHAL RADIAL h I

_ t i _

turERIAL VALLEY fk g

f l -. 1 4 PARKFIELD q -

)

g fi 3

6 SAN ONOFRE # *

. ~

}

d 4

7:

SOUTHERN b CALIFORNIA - #- - =

I coast I I lj u

3 D

e The height of the brackets correspond to equal absolute magnitudes for each  ;

trace. Hence, a large bracket corresponds to o small arrival, i FIGURE 4-12 COMPARISON OF GREEN'S FUNCTIONS FOR THE FOUR GEOLOGIC MODELS fy TERA-DELTA. 49 FINAL REPORT d

_h 4 osvisio= o+ rama commaario=_.)

t{-10 N

.- i i

SITE SPECIFIC RESPONSE j I PARAMETER STUDY  :

t EFFECT OF GEOLO'GIC L STRUCTURE 5

'1 MAXIMUM HORIZONTAL COMPONENTS EARTHOUAKE D 2 7 -

,w -"

']

.~_. - .- .

• 2'O 4

l

-%^ywyygn88

.f i N Mh_a ;'c/ /Ni NrNN/Af% /x: ( AV4,%/ W /N1 i V/ A A NM# %.O

. 2 l

I/Y A , k /bY(I b / b W

/[ >

M hNk

& Yklh%&iWW vy XN V 4' JN N A A ////4 9"xM 45 / ~ -A

'/ N .r 4 A nN i v /i ^

[ N[/Nk j d

'(

A:. VsN//V e gN P'd'Y)O / I'N!

, /< _

s

?  %

,, X'N sjr oM s

i +

' Dk'f y el Xi

,y - s.n o 2rr.

gmhD

,y, / ,, _ _ _ s c.i.,.,n c.... >1\ . /

x%sg= ^j '

7y. _ . _ e.,u .,4 i,nperi. v. .y ,/,(

f. y i

,, 'iM/ / ,1

(/ / s N'

\. e',

\,

s 3 g g-' r n; rg -r 3 ty p q q-~

'Ye$46. -.

N Hypothetical Smoothed Two Percent Horizontal Hesocnse Spectro that Would Be '

Found IllustratedWere in Figure the 4-!!.Son Onofre Site Located on Each of the Geologic Structures FIGURE 4-L

- I EFFECT OF CEOLOGIC STFNCTURE i ON SITE SPECIFIC RESP')NSE

_ __l-

_ A OfWisaO8e OF TEna ConPona F,Oas f

l 4-21

1

. . j i

I Magnitudes for the IV-79 event The value of My , based on the seismic moment, for IV-79 is 6.5 as given by Kanamori and Regen (1981). The 6

seismic moment is 6 x 10 dyne-cm, considerably smaller than estimates for the 1940 Imperial Valley earthquake.

(More recent results using IDA long period data give a moment 25 dyne-cm (Kanamori, sonal communication, value of 7 x 10 1981)). Following Slemmons (1977), the average fault length and displacement inferred by Kanamori and Regen, 35 km and

~

57 cm, respectively, correspond to M = 6.44 using the equations relating LD to M (Curve E, Table 15). The average offset of about 1/2 meter, confirmed by geodetic observations (Snay, 1981), corresponds to M % 6.4 (Curve E, Table 11) . Thus, we can see that the combination.of average displacement and fault length correspond approxim$tely to an M = 6.5 event, in agreement with the Kan'amori and Regen value for My.

The Ms.value rep rted in the USGS EDR, 6.9, is 0.4 units higher than M,. However, this M value is unreliable s

and biased upward because of the heavy weighting of European stations (10 out of 13 stations used) at a narrow azimuth along a path where the attenuation is relatively low. If the European data are together taken as one determination of M s

(= 6,9) and averaged with the three other values, an M s Of 6.(5 is obtained. The IV-79 magnitude is further increased by about .2 units, because the EDR uses the Prague Formula rather than the Gutenberg-Richter formula, which was used for the determination of M s f r IV-1940 (Kanamori, personal communication, 1981). Direct comparison of amplitudes recorded at DeBilt and Stuttgart 1"ndicates that the IV-40 l l

earthquake was larger in magnitude by .6 units (DeBilt) to .4 units (Stuttgart). Thus, this data confirms that the IV-40 event was about 1/2 magnitude unit higher than the IV-79 event. Considering the poor sampling resulting in the EDR M, value of 6.9, a value of M, and M y of 6.5 for IV-79 seems more reliable and consistent with the observec displacement data, especially in relationship to the IV-40 earthquake with M, and My = 7.1. There is no doubt that the displacement was much less for IV-79 than IV-40, and the corresponding M s should also be less.

The local magnitude, M3 , for IV-79 reported by Pasadena, is 6.6. Calculations bcsed on synthetic Wood-Anderson seismograph responses from strong motion records give M3 s 6.3 from Mexico stations (Brune et al., 1981) and ML s 6.2 from United States stations (Kanamori and Regen, 1981). The local magnitude for the IV-40 event ranges in magnitude from 6.3 to 6.5 (Kanamori and Regen, 1981). Kanamori and Regen note that, since for the 1940 event only statiens north of the border were available, there may be strong effects of rupture propagation. Since the 1940 event is believed to have ruptured to the southeast (Trifunac and Brune, 1970), away from the stations used to determine M , and away from the El Centro strong motion 3

station, its true local magnitude may have been considerably higher (because of defocussing of energy in the direction of the U.S. stations).

Estimates of the maximum Modified Mercalli intensities in the near field of the IV-79 earthquake are less than corresponding intensities for the 1940 earthquake, about IX vs. VII (Reagor et al., 1981). This is, in part, because of the longer duration of shaking for the 1940 event (in turn associate'd with the longer f ault length and greater fault slip). However, since there is a general correaltion of-peak acceleration with intensity, it is probable that peck accelerations in the 1940 event, particularly to the soutehast

~

nearer the large displacements and in the directior. of rupture propagation, were considerably higher in 1940 than in 1979.

9 l

l 1

o 9

L N s or M w versus M '

! L .

8 l -

e Kern County ,

i 7 Borrego Mt.. . le .i Irnperial Valley v )

.4 San , .

(1979) Long N M, rancisco Beach O ,

%. Guatemala 6 --

/

/ '

Imperial Valley (1940)

< San Fernando

/ '

Parkfield I

I I

5 5 6 7 8 9 NS rM g Fig. 4 KANAMORI & REGAN

Peak horizontal accelerations for IV-79 There are a range of ways of considering the peak acceleration data from the IV-79 earthquake. On the one hand, one might make the reasonable assumption that the high acceleration at Bonds Corner (.81 g) was the result of breaking of a local high stress drop asperity at some point along the fault, or the result of directivity due to rupture propagation (IIartzell, personal communication), or some other at present not understood effect, and take this observation as the most conservative point from which to extrapolate a maximal design acceleration for SONGS.

Assuming an M y for IV-79 of 6.5 (Kanamori and Regen, 1981) and a Design Base My for San Onofre of 7.0, along with the peak acceleration vs. magnitude d,ependence of abou,t .3M given in USGS Open File Report 81-365 (and Hanks and McGuire, 1981) yields a peak acceleration of about 1.1 g.

If a further correction for the lower attenuation at the SONGS site is compared to the attenuation in the Imperial Valley is made, s 1.8, (TERA-DELTA , Final Report, Fig. 4-13) ,

a peak acceleration at SONGS of s 2 g is obtained. This sort of extremal analysis indicates, as do many other consideration, that accelerations over 1 g and up to 2 g are possible. Of course, for a hypothesized M = 7 on the OZD, such high accelerations are less likely than lower values, and hence if we are willing to be less conservative, we can reduce these values in a number of more or less arbitrary ways:

l

l l

(1) We can assume that a " saturation" effect reduces the increase in acceleration between M = 6.5 and 7.5 to an average of only s .15 g, or even none at all (least conservative assumption).

(3) We can assume that the increase in peak acceleration between SONGS and IV-79 due to different attenuation (Q) is overestimated in the TERA-DELTA studies and reduce the correction to a factor of about 1.3 rather than 1.8 (FSAR Question 361.55 assumes a factor of 1.3 -

1.35), or even to 1.

(3) We can assume that for some as yet unestablished reason, the Bonds Corner record is unreliable and ignore it.

For example, we might take .61 g at Station #4 seven km from the fault, as the basis for extrapolation.

(4) We may take average accelerations recorded at -

various stations close to the fault extending arbitrary ,

distances along the fault away from the zone of maximum surface displacement, e.g., averages extending to the northern end of the Brawley fault and to the southern end of the Imperial fault, or some lesser or greater distance. This implies probabilistic considerations based on the fact that it is less likely that SONGS would happen to be situated at the

" worse" location than at some more " average" location. The

" average" can be reduced arbitrarily by extending the average to include stations farther and farther from Bonds Corner.

(5) We may consider probabilistic distribution and estimate the mean and one standard deviation in various ways, and assume that a mean plus one standard deviation value is sufficiently conservative.

The above considerations allow us a wide range of estimates of less conservative values for a DBE based on the IV-79 data, ranging from a value for peak acceleration anywhere between over 1 g and less than .5 g. However, these estimates cannot be rigorously justified from solely scientific considerations, there is some uncertain risk implied in the various assumptions made to obtain a given number.

e e'

e 9

Correlations of Peak Horizontal Acceleration with Macnitude and Distance The most recent USGS correlation of peak horizontal acceleration and velocity from strong motion records, in-cluding records from the 1979 Imperial Valley, California earthquake, are given in USGS Open File Report 81-365 (Joyner, e t al. ,1981) . This correlation is based on moment

magnitude, M3 , which is closely related to Ms* PriO#

• studies made correlations with M 3 (Joyner, et al.,

1981; USGS Open File Report 80-115), and still earlier a correlation was given without the IV-79 data (Circular 795, Boore, et al. ,

1978). In all cases, the data base is very limited, especially for large earthquakes at close distances, and the results must be " treated with caution" . .

The data base and understanding of ground motion is not suf ficient to place much confidence in these correlations, or other such correlations based on our present data set.

Reasonable assumptions in choice, weighting and elimination of data, in choice o; curve shapes to fit to the data, and in ways of grouping the data can lead to quite different results. Thus, as might be expected, different persons can obtain significantly different estimations of the peak acceleration for an Ms = 7.0 event at 10 km. For example, TERA - Technical Report 80=1 (Campbell, 1980) obtained different results than USGS81-365. Similarly, the WCC I

l

" Empirical Approach" results referred to in the SER give still different results.

The value from the TERA report for the 50% value (mean) for M = 7.0 at 10 km is considerably less than the corresponding USGS values. The exact reasons for this are uncertain at present (Boore, personal communication, 1981).

Note that one effect of the IV-79 data has been in part to lower the mean and one standard deviation curves for M = 6.5 (USGS Circular 795 vs. USGS81-365 results) . Among other things, this could be a result of the unusually high attenuation in the sediments of the Imperial Valley.

Data base and standard deviation Beca'use the IV-79 aarthquake is the only earthquake of magnitude gerater than 6 for which there are such a large number of stativ?s within 10 km of the rupture surface, there is a tendency to rely heavily upon its data. However, we have no basis for saying it is typical or that its accelera-tions are typical. In particular, there is no basis for assuming that the standard deviations estimated on the basis of our present data set represents the standard deviation of a population of different earthquakes of magnitude near 7.0 at distances near 10 km. At close distances we may find the standard deviation for a population of data from numerous earthquakes with different stress drop and rupture

^

characteristics, to be considerably greater than the standard deviation of multiple observations of a single earthquake, or of our present data set for only a few earthquakes, especially when most of the data is at larger distances. This could be particularly true if large earthquakes are characterized by a complex rupture process with large, high stress drop asperi Les, as many recent studies suggest. For this reason, I feel that the actual standard deviations could end up considerably greater than present estimates Magnitude " saturation" Magnitude saturation is often based on the assumption that the slip, stress drop, and energy release on a parti-cular section of the fault near a site, assumed to control the strong motion, will not change with magnitude, i.e.

that larger magnitudes will be associated only with longer rupture lengths and that the additional energy release from distant parts of the fault will not significantly change the strong motion near the site. However, it is well known that the amount of displacement on a fault increases with fault length, 2nd magnitude, up to magnitudes greater than 7.5. For example, Slemmons' (1977) compilation of North American date indicates maximum surface fault slip i \

\

t

l '

i l

I 23 8 8 i 1

1

.

O THRUST O i 22 - STRIKE- SLIP o O D

O O

0

'sO O[

ai -

• d _

"E u -

EJ 20 /

/

A g/ _

oi . [, / A' 3 ,9,/ &b '

19 -

g D

k/ h

/ A e -

l

/$ /no l

/ / ,,N 18 j /d

/

/

/

/

[7 ' I , , ,

26 27 28 29 3o Log Ma , dyne-cm Figure 1

. l I

i 1

j SCHOLZ, 1981

of .66 m for an M = 6.5 earthwuake, 1.71 m for an M = 7.0 earthquake, and 5.74 m for an M = 7.5 earthquake (Table 11, Curve E). Scholz (1981) (attached) has cast this result in terms of stress drop and fault slip as a function of fault length and moment (for strike slip earthquakes; 25 s 10 bars.for a moment of 2 x 10 dyne-cm, and s 100 bars 27 for a moment of 4 x 10 dyne-cm). (See Scholz, Figure 1.)

This data suggests that the amount of energy released on a gisen section of a fault (e.g. , + 20 km from a given point) may have a clear increase with magnitude. Thus, the amount of seismic energy a structure such as San Onofre would be exposed to would increase with magnitude. However, the rate

't a which peak acceleration would increase with magnitude depends on the details and' coherence of the pattern of, energy release. ,

As Scholz (1981) has pointed out, the dependence of high frequency energy release per unit fault length with increasing fault length (and magnitude) depends on the mechanism causing the increase in fault displacement with fault length (and magnitude) and this is not understood at present. Thus, one cannot confidently assume that " magnitude saturation" of peak ground acceleraticn will occur at some magnitude below 7.5, although it could begin to be effective at M = 6.5 or M = 7.0. It is even possible, given the uncertainty about the mechanism responsible for the large increase in average slip between M = 6.5 and M = 7.5, i

that the peak accelerations could increase somewhat faster with increasing magnitude between 6.5 and 7.5 than between 5.5 and 6.5, i.e., even faster than the Hanks and McGuire (1981) and Joyner et al., (1981) results.

Extrapolations based on the Recent Hanks and McGuire Results A recent Hanks and McGuire (1981) study has superseded the previous Hanks and Johnson (1976) study, often cited in discussions of magnitude saturation. Hanks and McGuire (1981) studied more than 300 horizontal components of ground accelera-tion from recent earthquakes and obtained the following result for the increase in average peak acceleration with magnitude: log a max N 0.30 M for 4 1 M=M g I 6.5, " remarkably close to that recently determined empirically by Joyner, et al.,

(1981) for 5.0 1 M I 7.7, their coefficient on M (moment magnitude) being 0.28 + 0.04." Thus, based on their data, in the magnitude range 5.5 to 6.5 there is no longer any indication of " saturation."

The Hanks and McGuire curve for peak acceleration at R = 10 km (distance to fault s 6.8 km) is significantly higher than the USGS81-365 curve (but has aoproximately the same dependence on magnitude in the range 5.5 - 6.5).

These results are baasd on a reasonable physical model for peak and RMS accelerations and thus may not be subject to quite the same arbitrariness in choosing curve shapes as the

O i

. .. i i 1 . . .

i i .i.

E l x .

y __

I o3 _ _ _ _ _

I l

o-u a e  ; -N o '/g e

o A,r ,r - i -

c -JBP R=73 km

^ -

a v - ' *' -

MI R = 10 km-ca -

o a o - -Donovan a> ^ v -

m n

• R= 0 km x v ' e- -This study S R =10 km E 2 -

.3 R = 10 km o o s .

E i O _.

t I

O J

' 9 i l i I -

I 7 8 A 4, 5 '

6 e

Magnitude B

l C 1 .

~

1:

r 1

e the other correlations. The mean value for M = 6.5 at R = 10 km (distance s 6.8 km) is .50 g. If we assume a ,

i standard deviation factor of 1.5, a value for a mean plus one standard deviation horizontal acceleration for M = 6.5 is .75 g. Concerning extrapolations to higher magnitudes, Hanks and McGuire state that on the other handk, a max need not increase by much above s 1/2 g for M > 6.5 at close distances, and we expect it will not, the linear increase in log in a""* above M = 6.5 assumed the empirical relations of Donovan (1973) and Joyner et al., (1981) not withstanding.

In a recent personal communication, Hanks has explained that this was not meant to be a precise statement, and could include increases ranging between, on the one extreme, a continuous linear increase at the same slope as between 5.5 and 6.5 giving accelerations of about 0.9 g at M = 7.5 and R = 10, to on the other extreme, a corner and flattening at 1/2 g.

Thus, to correct for an M = 7.0 event we could, on the one hand, assume the same slope on the acceleration versus magnitude curve between M = 7.0 and 6.0 as between M = 5.5 and 6.5, which would yield a mean acceleration of about .72 g and a mean plus one standard deviation acceleration of 1.06 g at M = 7.0, or on the other hand, we could assume complete flattening of the curve or complete " saturation" at M = 6.5 giving the same value at M = 7.0 as at 6.5 (.50 g). Assuming such complete saturation is not reasonable and would be the least conservative extreme. A more reasonable assumption is is that the curve has a decreasing slope between M = 6.5 and M = 7 giving a value of about .6 g for the mean acceleration a..d a value of .90 g for the mean plus one standard deviation acceleration. For an M = 7.5 event, the expected accelerations would be still higher. Thus, the recent Hanks and McGuire resulti suggest peak horizontal accelerations could be considerabi;r above the values quoted in the SER and also above the design base acceleration of 0.67 g for SONGS. The question of magnitude saturation cannot be solved by debate over the present data set, but must await accumulation of more data.

The fact that the magnitude saturation originally hypothesized by Hanks and Johnson (1976) for 5.5 5 M 1 6.5 has turned out to be incorrect suggests that we should not, without further verification, assume it applies between 6.5 M I 7.5.

Note that if the design magnitude is increased from 7.0 to 7.5, still higher accelerations would be expected.

e i

i 1

Conclusion regarding correlations of peak horizontal acceleration The recent USGS Open File Report 81-365 represents an extrapolation based on one set of assumptions, but as the authors point out, "for distances less than 40 km from earthquakes with M greater than 6.6 the prediction equations are not constrained by data and the results should be treated with caution. " The same applies to other attempted extrapolations. Whether mean peak horizontal accelerations increase with mangitude above M s

= 6.5 at the same rate

~~

they do below M, = 6.5 depends on how the effects of increased energy release per unit fault length balance the near field tendency toward saturation due to fault size, and

. this is not known at present. Recent fesults of Hanks and McGuire give higher accelerations than USGS81-365.

The above uncertainties, differences, and changes in the various probabilistic correlations, are to be expected in view of the limited data base, and are a reflection of true uncertainties in making estimates. The uncertainties cannot be eliminated by debate about various correlation methods and data selection procedures, but must await the accumulation of new data.

~

_61-

Vertical Accelerations In a number of recent earthquakes, recorded vertical accelerations at close distances have been higher than corresponding horizontal accelerations, at variance with a common earlier assumption that vertical accelerations would be 2/3 the horizontal accelerations. The design vertical acceleration for SONGS is .44 g, 2/3 of the horizontal design acceleration of .67 g. Thus, the recent evidence that near faults in some situations vertical accelerations may be higher than horizontal accelerations is very significant.

It further emphasizes the lack of understanding of earthquake strong motion and the fact that each new earthquake can yield unexpected results. In this situation there may be a

, tendency to look for "special explanations" for each unexpectedly high recording, with no corresponding effort et look for "special explanations" for unusually low recordings, thus biasing the average data downward.

In the case of the high vertical accelerations recorded in the IV-79 earthquake, several explanations have be.n proposed. Because of the lack of data, we cannot be sure that such high vertical accelerations are unusual. Among the special explanations suggested for the IV-79 high vertical accelerations are special constructive interference in wave guides (Archuleta, 1981) and local site amplification (Mueller and Boore, 1981). The high vertical accelerations recorded for the Gazli, Russia earthquake are reported to have resulted "becauso the fault beneath the site ruptured L

vertically upward towards the site (Hartzell, 1980)." (SER,

p. 2.5 - 33). However, no rigorous'modelling has been done to demonstrate this explanation. The Coyote Lake earthquake is reported to have "resulted in high vertical acceleration at one station because of S to P conversion at the interface between the soft alluvium and firm bedrock at depth" (Angstman et al., 1979). Other cases in which vertical accelerations have been reported higher than horizontal include the Naghan, Iran earthquake of April 6,_1977 (Ambrayseys, 1978) and the Victoria, Baja California earthquake of June 9 1980 (Simons

'et al., 1981). Perhaps the most important example is the 1933 Long Beach earthquake, since it occurred on the OZD.

Ratios of vertical to h6rizontal peak acceleration were 1.45, 1.00 and 0.60 at stations 6, 9, and 12 km from the ,

fault (Luco, personal communication) .

At the present tima, we do not have a certain enough understanding of vertical ground motion to be sure that, at distances less than 10 km, high vertical accelerations will not be common. As one example, a recent model by Blandford (1975) attributed high P wave excitation to asperities on the fault surface. Models with asperities have been suggested in numerous recent studies. If any similar mechanism were operating in many earthquakes, then close to faults, vertical e.ccelerations higher than corresponding horizontal accelerations I

could be quite common. Given our present lack of understanding, I and lack of data, we cannot be sure of the degree of conservatism involved in the vertical design acceleration of 0.44 g.

It has already been exceeded in several earthquakes and there is no reason to believe it could not be exceeded during an earthquake on the OZD.

Special mention needs to be made of the explanation for the high vertical accelerations o'bserved in IV-79 given by Frazier in his testimony regarding the Diablo Canyon Nuclear facility. As Dr. Frazier pointed out, "at close distance, the peak horizontal accelerations were as high, on the average, as the peak vertical accelerations."

(Written testimony of Gerald Frazier, p. IV-1.) Dr. Frazier then goes on to explain the characteristics of P waves and S waves which relate to the ratic of vertical to horizontal accelerations. He describes the " echo chamber" effect due'

~

, to the high near surface velocity gradient which "would tend, to' increase both P and S wave amplitudes...However, high frequency S waves are severely attenuated in the shallow sediments, thereby compensating for the amplification of the sedimentary basin. P waves, on the other hand, can travel efficiently in the soft surface materials. In the shallow sediments in the Imperial Valley, S waves above 10 dz can be attenuated by a factor of ten within one kilometer, while P waves above 10 Hz are attenuated only about 20% j over the same distance." (Written testimony of Gerald Frazier, Diablo Canyon transcript, p. IV-4.) Thus, in this 1

explanation proposed by Dr. Frazier, it is suggested that the high vertical relative to horizontal accelerations occur because the S waves have been severely attenuated by the low Q for shear waves. It follows that if an earthquake with the same stress drop.had occurred in a structure with higher Q, but approximately the same shear wave structure, the peak horizontal accelerations could have been considerably higher. This would be an extreme case of higher accelerations on rock vs. soil sites.

Thus, reduced horizontal accelerations'for the IV-79 earthquake may have occurred as a result of high material attenuation (low 0) in the sediments near the earthquake, especially for higher ho,rizontal accelerations observed close in, where non-linear effects could result in the most attenuati on. This possibility further indicates we should be cautious in. assuming the peak accelerations observed in the IV.-79 earthquake are typical.

e

Ground Motion Implications of Northern Baja California Earthquakes Since the northern Baja California seismicity near the San Migtel fault zone is in a tectonic environment similar to that of the OZD (part of the same linear zone parallel to the plate boundary, recent activity , a smaller amount of total slip, probably in similar deep basement rocks) it is reasonable to assume that an earthquake similar to, but larger than, the 1968 El Alamo earthquake (M = 6. 8 ) could occur on the OZD. .

. This raises a question that has not been adequately dealt with in either the FSAR or the SER. Because of the special characteristics of earthquakes in the region, i.e.

relatively high inferred stress drops, an earthquake with the same M, as an earthquake in the Imperial Valley (e.g.

IV-79 or IV-40) c'ould generate much greater high frequency energy. A number of studies describe the characteristics of these earthquakes. Brune, o c a l. , (1963) noticed a region of anomalously low surface wave excitation for a given M L ( r, conversely, greater high frequency generation for a given Ms ) in n rthern Baja California while studying events from the California-Nevada area. A similar result, expressed in terms of apparent stress, was found by Wyss

~

and Brune (1971). Typical values for apparent stress para-meter defined by them were 140-710 bars in the region of the San Miguel Fault Zone, as compared to typical values from less than 10 up to 100 bars in the Salton Trough region.

High apparent stresses were also observed in certain areas

. of southern California. Thatcher (1972) studied the regional variations of spectral parameters in northern Baja California

~ from observations at Cal Tech stations in southern California especially BAR, PLM and PAS. The spectra were interpreted in terms of the source parameters moment, source dimension, stress drop and high frequency spectral fall-off. Thatcher inferred that northern Baja sources have dimensions that are typically a factor of four smaller than the dimensions of

. Gulf events of comparable local magnitude. Conversely, moments.for the Gulf events were about an order of magnitude larger (for the same Mg ) than those for northern Baja. The average stress drop for the Gulf earthquakes was found to be lower than the average for northern Baja California. Nava and Brune (1981) and Nava (1980) did a more detailed study of a sample earthquake from each region, the Pino Solo earth-quake from northern Baja California and the Mesa de Andrade earthquake from the Colorado Delta. Both earthquakes have 4

approximately the same local magnitude (determined from the maximum amplitude measured on standard Wood-Anderson short period torsion seismographs). However, the relative long

period excitation was much larger for the Mesa de Andrade earthquake.

To obtain a more quantitative comparison of the relative high frequency excitation of northern Baja events, we may compare the various figures and spectra shown by Thatcher (1972). The spectral amplitudes at 4 Hz are typically a factor of 10 or more higher for.

northern Baja events than for West Coast of Baja or Gulf of California events of about the same moment. We can also estimate this difference from Table 2 of Thatcher by extrapolation from the corner frequency using the tabu-lated slopes of the spectra. Again, we find that the amplitudes estimated at 4 Hz are typically 10 times higher for the hiigh stress drop northern Baja, events as compared to the Gulf of California events o,f about the same M (moment magnitude).

Since peak accelerations would be expected to be closely related to the excitation of 4 Hz energy, the above results are a clear indication that the near source peak accelerations for the norther Baja type high stress drop events can be considerably higher than for events of comrarable M, in the Imperial Valley. The actual difference in peak acceleration will depend on the nature of the rupture mechanism. Nava (1980) and Nava and Brune (1981) inferred that, in addition to the factors cited by Thatcher (1972) for northern Baja events, namely, lower source i dimensions (for a given Mg ),.and higher stress drops, I

s. .

l . .

l .

l

( another factor operating in some cases was the relative simplicity of rupture for the northern Baja events. Rela-tively lower average stress drops in the Imperial Valley might be expected on the basis of the high heat flow, thick sediments and high water content. Whatever the physical mechanics responsible for the differences between the j

)

Salton Trough type earthquakes and the northern Baja type earthquakes, the direct observation of greater high fre-quency excitation for northern Baja events indicates that l

assuming that the accelerations observed for the IV-79 l l

earthquake are " typical" or " conservative" is unwarranted, and earthquakes of the same M, can have considerably higher near scurce accelerations.

O

Implications of the El-Alamo Earthquake of 1956 The 1956 El Alamo earthquake has been considered in FSAR Question 361.68. However, in that consideration, it was assumed to have the same characteristics as the 1979 Imperial Valley earthquake. In view of the above cited evidence for differences between northern Baja earthquakes and Salton Trough earthquakes, I do not believe the 1956 El Alamo ea.thquake has been adequately considered.

There is strong evidence that the El Alamo earthquake generated higher accelerations than the IV-79 earthquake .

at comparable distances. For example, the intensity map for the El Alamo earthquake shows an area of intensity VI about thirty times larger thar. the area of intensity VI for IV-79 (compare attached figures).

The high amplitude and high frequency content of the record at El Centro is in striking contrast to the records at similar distances from the IV-79 earthquake. It is mentioned in the FSAR Question 361.68 that the rupture in the El Alamo earthquake was to the southeast, away from San Diego. If the rupture had proceeded northwestward directively focussing of energy could have led to even higher intensities in southern California. Thus, available evidence indicates that the El Alamo earthquake had

.4 s.

A >5 , - h .

5.

a J a .

~.

e d l *

• i

.l 1,

l l

! ~

\ l .

&j

.i 1.- /

-- ;_ 7  :

~

j f-

/ .,

~

m  : . g -

<1 o

[

i 8 .

l

/g ~

y

+ + ..'

• w p! i

/ :g 2g-e

..? JI. '~>f R(/ m -

pe I 1,- <

a e

I bi j.

bl

.s. <

{ . 1

/

(! (!,q f !N_ g s

'f ..I 1 -  ;, .i

• e4 y

/

e l

• B ,

4 . \ .

e s i u1i

• ei 3 a i ; i n

};

,g -

/- 4 g

" ~. 3 3 y

.3e ,/

I/ se*

n

- e l *l i

i

] - .

. . a

-e ~

e < , i t

o  ::-  ! s' a

3 *

• s 3;

g. *>a ?g.

= s -

e

~

si 5 n- . / 5  ? *$ i p=i, ,

3 atsa 1 =a c332 g .- -

.1 8 5#- -

a g , ,. . 3 g

.g

. .I i.

(

.1 3o

• 3 .

s ...

I J

< a ~

-.5,~~

=

~

• ; r . -

-a 2

.8 - -

c

.i - ec 2 . . T.. .,

a .

~

1 o 2 5 s ..

3, .- . ...

, ' =

. , 3 a .>

-.4 g

4 J

a v - . , . . ,, #

o

. .  : w. . / .

+ ..

• ",."l..', .*

• .EE

• g . * ~ ' . *  !

- . K' 1

. s. *$ f l

'  % \c .

Q k-

.. ,Y

= .

,/ ..

• _e

~

REAGOR, ET AL. and g g U.S. EARTHQUAKES, 15 40 k 3 , J w (Inset) h a .

120* to IIS* 118*

lil' 116* 115' 114'  !!3' 112'

'\, NEVAbA .

\ $'

. .O

. calif 0RNIA , O \, j p . . .

35* -

O \./ 35' O

O O O '[O4 O Needles e

0 0 4'

>4 0 0 0 0 O 00 4 g*gy o ARIZONA o o .

7 4 o3 3 3 4 4 4 \. U 4 04 'O 30

' 30 4 4 5 4 08 jj' --

O 04 4 05 O O /gO

$ 4,[ 5 gedes M ugg og P . _ .

4 gos A 4 04 0

SI o

S

%g O

• O y o y 4 O M M

l 4

u. -

'tWe Sen D go 6 O

66 5 0 #

• VI O

• 5 O

6 0g g [,,,,o %g t

s o

'-~

s....

O 90 30

,,. l n

u 6

o O O

,2 -.? 8 ~~~

~~6 ,,9 6

-- O O 4 4 l .'e.

-(**bf,6ym,y s *

.C

, ) .

tn 32 LOWl'R Calif 0FNlA F 's '

~~ ^

32' February 9,1956 N-O 0 S0

- >--, w ,100 .

Statuto Males 0 **

t; g IF 119* 118' ilP 116* 115* 114* In*

> ur M

M Na as 8 - Area aftertel by earthquake et Fetsruary 9.

in W

m

?,

% [$3 hO k h[ h hh -

significantly higher near source accelerations than the IV-79 earthquake. A thorough study of-the records of

. this earthquake should-be made, rather-than an unwarranted assumption that it was similar ' to IV-79.

i a

1 e

e

.71-

.1

Intensity Magnitude for the IV-79, IV-40 and El Alamo 1956 Earthquakes We may use the area of intensity VI to define an Intensity Magnitude, M7 , using the relationship.of Hanks.,

et al. (1975) between area of intensity VI and mcment:

Log M g = 1.97 Log Ayy - 2.55 and the Hanks and Kanamori (1979) definition of Moment Magnitude:

Mg =flogMg - 10.7 Combining these relationships we define Intensity Magnitude as:

My = 1.31 log Ay7 -

12.4 Using this result we obtain, for the 1956 El Alamo earthquake M

7= %7.3 (Ayy% 10)1 cm ), for the IV-40 earthquake, M7 = 6.6 (Hanks, et al.,1975) and for the IV-79 earthquake, M 7= 5.3 (A77 N3 x 10 13 2 cm ) . This dramatically illustrates that the intensity of shaking, damage and high frequency energy in southern California was very low from the IV-79 earthquake and thus, the use of IV-79 peak accelerations as typical of southern California earthquakes is unjustified, and along with other evidence cited above, indicates that the assumption made in FSAR Question 361.68, namely. that the accelerations for the 1956 El Alamo earthquake were the same as for the IV-i 79 is also unjustified. )

i l

\

l

m . - -

0. a -

New Strong Motion Data Victoria strong motion record of the June 9, 1980 northern Baja California earthquake (See attached Abstract, Simons, et al., 1981)

This record 2 ig for an M = 6.2, M = 6.3 event had t s peak vertical accelerations of over 1 g and horizontal accelerations of up to .85 g. It is,of particular interest because it occurred in an environment similar to that of the IV-79 earthquake and, thus, indicates that peak horizor.tal accelerations similar to that recorded at the Bonds Corner station (.81 g), and vertical accelerations of over 1 g, are not unexpected for this magnitude (M N 6h) earthquake in the Imperial valley. No surface break was ob' served along the primary fault (the Cerro Prieto fault) and at the present time, no fault plane solution or knowledge of rupture parameters is available. It is possible that the horizontal projection of the rupture at depth passed within less than 1 km of the station.

Previous earthquakes -in this region, of magnitude less than 5, generated peak accelerations up to about .6 g at the same station and, thus, there may be local site conditions partially responsible for the high accelerations. Both I

, _ _ _ t-

vertical and horizontal response spectra for this earthquake exceed the DBE for SONGS 2 and 3, at damping of 7%. A final assessment of the Dnportance of this recording must await a better understanding of the rupture mechanism of the earthquake and better knowledge of the local structure.

Mammoth earthquake accelerations (John Anderson, personal commun. cation, 1981)

The May 27, 1980 Mammoth earthquake, M L = 6.2 (BRK) 6.3 (CIT) , M, = 6. 0 (GS) , Mb = 5.7 (GS) was recorded on three temporary strong- motion stations at distances of approximately 10 km from the epicenter (See Map 1 attached) .

One station, " Convict Lake", located near the outflcw of Convict Lake, recorded a peak horizontal acceleration of 0.72 g (See attached record) . The other two stations, located at McGee Creek Inn and Long 'e' ley Fire Station, recorded peak ac'celerations of 0.20 g and 0.35 g. Tha high acceleration recorded at the C-avict Lake station may have in part been influenced by directivity, but a reliable interpretation of this data point must await a better understanding of the earthquake mechanism. The station was not located on any obvious topographic or structural feature which could explain the higher accelera-tion recorded. The instrument was resting on the ground, not i

anchored to a pier, and this may have led to some distortion of the signal, i

l

-74

s .

i i e0822 40' -

+ i -

h 1480 ofg %y 3 9S

,3679

. Lake Crowley 3754 Convic t

+ \ 1494  %

35' -

Lake \Qdh +G .

, g\ 1454 -

Long '

vine,tation FirtS o1525 l l

1 rm s N Epicenter

~~ ~~ _ y 30 _

Ma y 27,1980 , _

1450 GMT I i 1 I I i O 5 km t i 50' 118 45'

. MAP 1

A 4

CON VIC T LAKE M AY 27,1980 1450 GMT M A M MOTH LAK ES , CALI F.

O 2 I I 4 6 I I i .I 8

i i l 800 -

A g UP

.o =\p4,w,xe,hgh.;;;npff - l lj.jp , ~,, _m -

P V)

N E

-800 -

3 800 -

r A

-~ - - - - -__ A,..dkAA,ichsi 16 5 ,

< 'n*wyyvwyv~pql!$J4.. j pj p v . v y ,

4- . -

- ~ ' --

,I  ;

$ -800 - .

d 8 0'0 -

o o d e

< .. 75

, _ A A

- y _vvpvv,..auhkN151 A.

~wy q -

~ -

i 1;y -

-800 -

~L I I I I d 2 I I I I 4 6 8 i TIME (sec) 9 O

v -

Ground Motion from Possible Slip on the Cristianitos

- Fault At the time of writing of this testimony, it has not, to my knowledge, been finally established whether or not the Cristianitos fault is to be judged capable. If it is, the distance to the fault . trace would be reduced considerably.

~

The expected peak ground motion would be increased. I believe that if the Cristianitos fault is judged capable, then the whole ' issue of ground motion must be re-opened, as the present FSAR and SER do not deal with this possibility.

e eme=

- 75-m- %t ~~'

=

IAICA A85 TRACTS. 76TH AMNUAL MEETIf4G 85

. the Imcerial Valle/ above 11 Hz. .The efface is frecuency dependene. and ioes

. ss and arrival times of not show a significanc reduccion for frequene:es below 5 Hz.

and origin times at tne For this particul:r situacion, at a diacance of 'abouc 3 k:n from the rupture surface of the faulc. the reduction is f be explained by usin9 more Pronounced for horizontal cocl onents chan for vertical.

entro Array. Within

• aired to (splein tha ts. Other investigations AN EMPIRICAL ANALYSIS OF THE SOURCE OF ENERGY RELEASE DURING THE ttion data on this CCTOBER 15, 1979 IMPERIAL VALLEY EARTHOUAKE CAMPBELL. KENNETH W. and MARK W. POLIT. TERA Corporation. 2150 Shattuck Avence. Berkeley. CA 94704 The extensive set of strong-motion data recorced curing the Octetier 15

' RAY STATION .i 1979 Imperial Valley Earthouake has been used to study statistically the probable sources of energy release en tne fault rupture surface re-Survey, Menlo Park' sponsible for producing tne pesK grouno motions at these stations.

t Gff;ct is resconsible This was studied by regressing peak grouno motion parameters on ois-tance for a series of potnt sources, iccateo both at tne surface and at t statioa 86 of the USGS depth. distributed at equal intervals along the ruptura from the ept-1979 Imperial Valley center to the northern end of the Imperial ano Brawley Faults. Alse

.tia greater than investigated was the effect of considering multiple point sources.

ons. Smoothed scectra . These results were compared to a similar expression wnten was baseo on e 5 ance % eacn station M me fault ruotm s# ace.

k. fa r t ee at eight Hz. 'atios Preliminary results for peak acceleration (PGA) indicate that no single aking duplicate the point-source model can reoresent the atteauatier. characteristics of tved linearly curing th> s parameter as well as the expression baseo on closest distance, at rec rds at station i Fur . :rmore, these point-scurce models recuire a zone of constant accel-hans owing to higher eration within aoout one nalf a fault-lengtn distance from the source d Brawley faults. A in order to best fit the cata. The point-scurce found te give tne he same stations con. least scatter and best fit of all single scurces investigated was 10 temporary statiens '

cated about 6 km. north of the epicenter. By eliminating cata ortrar-southwest of station ily south of the epicenter. tMs point was fcund to mig-ate scme 15 to Ith a peak near 12 Hz. 20 lei. north near the junction of the Imperial and Brawley Faults. 3y ral thinning of the considering two potnt sources.*a mocal almost as gcod as that for clos-Catton 86. est distance was obtained. ThSse results suggest that PGA data cannot support a single concentrated zcne of energy release but aather are consistent with either multiple Icnes or a single broac zona of energy

.RRAY TOR THE 1979

'crsity of Uash. . J iz. m ra- - J, THE GUADAltJPE VICTORIA ST9CNG FOTION RECORD FRCM THE NEAR FIELD OF THE

' ' JUNE 9.1980 NORTHERN SAJA CALIFORNIA EARTHOUAKE Kn applied to the SIM)NS. R. S., SRUNE. J. N. . ANCERSON. J. G., Institute of Geconysics

.o 1979 !ccerial and Planetary Physics. Scri os Institution of Oceanograohy. University r y evarages were of Califorria. San Diego. La Jolla. California 92C93; PRINCE. J..

hrr:nco of ground MENA. E.. Institute of Engineerinn. Universidad Nacional Autoncea de th:t miPhc be h Mexico. Apartado Pcstal 70-a72. Mexico 20, O. F., Mexico n citernative, 5.2 ncy binds i.'ere The June was Cal Tech) 9.1980. " Victoria" recorded earthouake (N.

on a strong-motion Baja California, instrument (1 g digitag =l Kine-e n staciens , metrics DSA-1) in the town of Guadalupe Victoria, sll km W of the pre-r:duccion factors liminary epicenter and close to the Ceero Prieto fault. The town of r:duccions of 10- Guadalupe Victoria exnerienceo severe drage, carth uahe for The early part of the recording suffers from low magnetization level and d fer fracuencies consequent bit confusion. A combination of hardware and software techrzues w

e

9 .

i 86 SE!5M0 LOGICAL SOCIETY OF M(RICA I

has successfully recovered mst of the data for the vertical and one nor-120ntal (N 40* W) component. The recovered data nave occasional gaps. nne l I

[

about .35 :ec long :nd the rest ! css tn:::the time scale. Fill feg the4 gaps accelerograms unicn jntegrate plausibly to velocity and ittsplacerent. {

Vertical accelerations reacn or erceed 1 o several times in cotn tne pos. <

itive and negative directions. In one 0.25 sec segment. tne instantaneous vertical acceleration exceeos 1 g 4 tires at frequencies as low as 10 H2.

The hor 12ontal acceleration reacnes 0.85 <t and exceeds 0.5 9 at time spanning an interval of -/l sec. Hort2cntal peak velocities are A0 cm/sec.

a ,

I j The Peergent beginnings of the records suggest a complicated source time history.

Acceleration response spectra nave been computed from both available

. components. using a variety of damping factors. In all cases, the horiz.

ontal resnonse spectrum eteeds tnat for tne Bond's Corner rec OctoJer 15, 1979 range from 0.0 to 0.3 sec. This recordino provides a unicue opoortunity to compare damane in the region (mainlyectrum adobe at and bri

. SVI!!. was surprisingly moderate in view of the hign response sp

,f short periods.

AUTHOR It0EX

. 59 sy suMGam u. 31 AsNEW. J. autotCA 2. J. 81 56 Ast. E. 23 90RE AU . G. J .

13

. ALAACoh. t. SUTLER. R.

ALLIANDER. s. s. so

• 44 ALGERMisstN. s. f. s tAGhtTTI R. s.

f AMINI. A. 21 CAMPSELL. A. w. 18. 19. 85 .

73 U ANDERs0N. J. G. 85. 37. 21. 17 CAPUTo. M. ,

12 CMytR. O. 62 '

ANDERsCN. F.

6o 27 ANGLlu. F. M. CASTRO. R. da APs[L. R. J. 79 CHAVEl. O.

62 53 ARABA12. W. J. Osm. E.-#.

I. DemG. 9. 4. 76 SACMt. f. C. 33 CLAAK. M. M.

45. 46 SAKUN W. H. 82 CLuFF. L. s. 3. 8. 70 i SENhdf. J. M. 28

• CLTE R . E . W , 72 SENNiff. f. J. 6o 41. Fo,11 CoPPERsMITM. s. J. 56 SERNRtufER. O. L. so. 76 C3ER.R.P.

to 28 , 4s i SE2CRt. s. P. CRAER. C. M. 16 BIRO O. N. 39 CUTLER. R. g 4

- SLUK. J. A. 53 33 SoLLINGER. G. A. OAGGtti. P. N. 18 l toLT. 8. A. 52,3333, 14 OARRAGM. R. 5. 64 5

toantR. 3. P. OAAT. R. L.

toGLER, J. R. ss 19 an. 34 CAvis. 3. J. F SooRE. O. M. OAVis. J. F.

toWWN.R. u OA91s. f. N. $7 SRA85 t. E. s s OtGENK0Lt. M. J. 41 SAAILt. t. W. 54 otNRING. N. E. i

f. M. $$ 64 '

laoCMtt OtWY. J. W.

lauht J. M. 23 1T.ff18. 45 000LAat. M. 20, 35 j BuCMANAN.8ANKs. J. M. ORAtlNsKl. M. 20,22 60

' BuCMSINCER. G. G. R.

64 Ost. C. M.

• Sutt. c. G.

i I

D N

l l

. l l

1 SCALING LAWS FOR LARGE EARTHQUAK?S CONSEQUENCES FOR PHYSICAL MODELS l

Christopher H. Scholz l Department of Geological Sciences and Lamont-Doherty Geological Observatory of Colusbia University, Palisades, New York 10964 ABSTRACT It is observed that the mean slip in large ear thquakes correlates linearly with fault length L and is not related to fault width, W. If we interpret this in terms of an elastic 'model, it implies that static stress jrop increases with aspect ratio (L/W). We als o observe a tendency, particularly for strike-slip ear thquake s , for aspect ratio, and hence static stress drop, to increase with seismic moment. Dynamic models of rupture of a rectangular fault in an elastic medium show that the final slip should be controlled by the fault width and scale with the dynamic stress drop. The only way these models can be reconciled with the oher-vetions is if dynamic stress drop correlates with fault length so that it is also nearly proportional to aspect ratio. This could only happen if fault length is determined by the dynamic stress drop. There are several serious objections to this, which lead us to suspect that chase models may I be poor representations of large earthquakes. Firs tly, it conflicts with the observations for small earthquakes (modeled as circular sources) that stress drop is nearly constant and independent of source radius. Secondly, l

l

3 2

it conflicts with the observation that fault length is of ten determined by rupture zones of previous earthquakes or tectonic complications. We speculate that the boundary condition at the base of the f ault, that slip is zero, is unrealistic because that ed ge is in a ductile region at the base of the seismogenic layer. In a modal in which slip is not so con-strained at the base of the fault nor at the top (the free surface), such that no healing wave originates from these ed ge s , final slip would be determined by fault length. The observations would then be interpreted as seaning that the static and dynamic deress drops of large earthquakas are nearly constant. These two alternatives predice very different scaling of the dynamics of large earthquakes. The width-dependent model predicts that average particle velocities are larger for long ruptures but the rise time will be the same as in a shorter event of the same width. The length-dependent model predicts the opposite. '

I e

. 4 - , _

- 3 INTRODUCrION A central problem in earthquake seismology has been to find scaling laws that relate the static pacewters such as slip and stress drop to the dimensions of the rupture and to understand these relationships in terms of the dynamic parameters, the most fundamental of which are rupture velocity and dynamic stress drop.

In doing so, it is essential to distinguish between small earthquakes and large earthquakes. Tectonic earthquakes nucleate and are bounded within a region of the earth between the surface and a depth h,, the seismogenic layer. The seismogenic depth, h,, depends on the tectonic environment but in a given region the maximum width of an earthquake occurring on a fault of dip 5 is W3 = h,/s in 6. We will define a small earthquake as one with a source radius r 1 W,/2 and a large earthquake as one in which r > W, /2. Thus a small earthquake can be represented as a -

circular source in an elastic mediuta, whereas a large earthquake is more s uit ab'.y created as a rectangular rupture with ons edge at the free surf ace .

It has been repeatedly demonstrated (e .g . , Aki, 1972; Thatcher and Hanks,1973; Hanks,1977) that the stress drops of small earthquakes are j

nearly constant and independent of source dimensions. This resul e, when interpreted with dynamic models of finite circular ruptures (Madariaga, 1976; Archuleta, 1976; Das , 1980 ), simply means that the dynamic stress drop is constant.

If the same were true f or large earthquak s , the dynamic models of rectangular faulting in an elastic medita (Day, 1979; Archuleta and Day,

+ = , s.

5 OBSERVATIONS For small earthquakes, using the definition of seismic moment, M, and the relationship Ao - $h where r is source radius, j is mean slip, and 60 is stress drop. If stress drop is constant, the relationship between M, and fault area, A, is M = (16Ao ) A3/2 . (1) o 3 p /2 Large earthquakes, however, are more nearly rectangular ruptures of vidth W and length L and in this case, for an elastic model in which slip is restricted to be within W, 40 = C'dW (2) where C is a gemetrical constant.

l If stress drop were constant, we would expect to find that l

Mo =kw2C (3)

In Figure 1 we show a plot of log LW n. log M,, for the largs interplace thrust and strike-slip earthquakes from the data set of Sykes

(

• a. ,

1 6 . -

and Quittaeyer (1981). These observations are ilsted in Table 1. The data for each type of earthquake define a line, but with a slope less than one, indicating that stress drop systematically increases with aanent. The offset betwaan the data for che, strike-slip and thrust events is also an important feature that we will discuss later.

These data indicate that I is not simply related to W and that ac is not constant for large earthquakes. On the contrary, many workers (e.g.,

Bonilla and Buchanan,1070; Stensnons,1977 ) have argued that u correlates with L, and recently Sykes and Quittmeyer (1981) have argued that the correlation is linear. Plots of u s v_s,. L on linear scales are shown in Figures 2 and 3 for strike-slip and thrust earthquakes, respectively.

In view of the usual uncertainties in the estimates of E and L, and any naturally occurring variations in dynamic stress drop (with which slip should be expected to scale), the correlation betwe,en u and L is fairly .

strong. We fit it with a straight line with an intercept at the origin u = at (4) and find that a e 2 x 10-5 for the thrust events and 1.25 x 10-5 for the strike-slip events. At least for the strike-slip events, slip is clearly not dependent on width because the widths of all the events in Figure 2 are between 10-15 km, i.e. , they are essencially the same.

From this observation we would then expect that '

P 2

l

+ -

Mo = us L w (5) l

7 which is confirmed in Figure 4. For reference, the line drawn through the data has a slope of one.

Since 2

b LW3A3/2( )

and since the aspect ratio L/W varies only by a f actor of about 20 in the data-set, we would have found a good correlacion between M, and A ! , as did Aki (1972) and Kanamori and Anderson (1975/ had we plotted log A m 2

log M,. The question is nst whether M, correlates better with L W than with A !2 The issue of concern is that Kanamoci and Anderson's inter-pretacion of their correlation as meaning that stress drop is constant is only true if L/W is con; tant, because from (2) and (4), we have L

aa = Cucg . (6)

That L/W is a constanc is an explicitly stated assumption of Aki (1967, 1972) and Kanamori and Anderson (1975), and although Abe (1975 ) and Celler (1976) accompted to observational 17 juscify this assumption, it is not generally true. In Figures 5 and 6 we plot ao vs,.

s L/W for the two types of earthquakas. The correlation between them is very clear for the strika-slip events, and less so for the threst events, for which there is a much smaller variacion of aspect ratio. That L/W does not have a large vari-acion for the thrust e*. cts seems to simply result from the fact that the seismogenic width of subduction zones , ,W , is about 100 km, so that only extremely large e' vents can achieve high values of aspect ratio.

We can now understand why stress drop increases systematically with M,, as shown in Figure 1. The width of large strike-slip earthquakes is

o . .

. e limited by the seismogenic depth to W ,

• 15 km so that they grow princi-1 pally in the L d! getion. This results in a systematic increase in L/W, and hence Aa , with d,. The subduction zone thrust earthquakes have dif-forent widths but L increases f aster than W with increasing moment, pro-ducing the same cesult, i.e., ao increases with L/W or M,. The offset between the data for thrust and strike-slip events in Figure 1 occurs simply because the widths of the thrust events are much greater than those of the strika-slip events. A strike-slip event must have a much greater as pect ratio, and hence stress drop, than a thrust event of the s ame moment

• e e

o e

e

.~,

9 PHYSICAL CONSEQUENCES The principal feature of the observations that we wish to explain is the correlation between slip and fault length. It is a surprising obser-vation because intuition would first lead one to expect slip to depend on width, yet this is not observed. This intuition is re-inforced by the results of dynamic models of rectangular faults in an elastic medium (Day, 1979; Archuleta and Day,1980; Das,1981 ). These models show that slip is controlled by the width of the f ault and that it scales with dynamic stress.

drop.

The situation is illustrated in Figure 7, which shows surface slip slons the fault for two representative s trike-slip earthquakes. These earthquakes have essentially the same width, and differ only in length. If the dynamic stress drop were the same for these two e ar thqua'kes , then according to t he theory, the Ft. Tej on earthquake would be the equivalent of six Mudurnu earthquakes placed end to end. Clearly that is not the cas e. -

If the dynamic, elastic models are correct representations of eare".

quakes, then the only way they can be reconciled with the observations is if dynamic stress drop correlates with aspect ratio. Since the width of strike-slip events is nearly constant, and the , width varies much less than length for the thrust events, this would be approximately true if dynamic stress drop correlates linearly with fault length. The only way this can happen without violating causalley is if fault length is determined by dynamic stress drop. This is not an entirely unphysical pro position ,

because dynamic stress drop determines the stress intensity f actor, which 1 --- , --m. . - . .. ., . .

10 ,

is important in fracture growth. It is not obviously apparent, however, why L should increase linearly with aed, the dynamic stress drop.

There are several major objections to this interpretation. The first is that we have to assume that for large earthquakes aaddetermines the rupture length, which directly contradicts the obs ervations for small earthquakes. Although se'ress drop appears to increase with source radius over a limited range in same data sets (Aki, 1980), it shows no obvious variation with source radius over a very broad range (Hanks , 19 77 ). We can offer no reasonable explanation for why large sarthquakes should behave different1y' than small earthquakes in this important res pe ct.

A sece objection is that th!s assumption conflicts with the prin-cipal observat:.as that led to the conce pt of seismic gaps: that the length of large earthquakes is of ten controlled by the rupture zones of previous earthquakes or by structural features transverse to the fault sone. Of course, one could sof ten the original assumption tot aa d deter- ,

mines the length unless the rupture encounters a rupture zone -f a previous earthquake or a transvsrse f eature. The rejoinder is that if the latter were as comanon as is, thought, it would have the effect o. destroying the correlation between u and L chat is observed.

It is worth giving a specific example. If we compare the 1966 Park-field earthquake (L = 30 km, u = 30 cm, W = 15 km) and the 1906 San Fran-cisco earthquake (L = 450 km, u = 450 cm, W = 10 ',an) we need to explain the difference in u by a difference in aa of about a f actor of 15. Since d

the correlation between u and L is also good in these examples, we also need to argue that aad determined L in chase cases. On the other hand, it can be argued that the length of the 1966 earthquake was determined by the length of the gap between the rupture zone of the 1857 earthquake (or the 9

. a f

, 11 fault offset near Cholane) and the southern end of the creeping section of the San Andreas f r.ul t . S imila rly , the 1906 earthquake filled the gap between the northern end of the creeping section at San Juan Bautista and the end of the fault at Cape Mendoc*no. If our .trgument' that aa d ***"

mines L is true, then these latter observations are coincidences. Almost identical arguments can be made for many of the other earthquakes in our data sec.

The third point is less an objection chan a surprising consequence of this interpretation. The Hoei earthquake of 1707 ruptured about 500 'ian of the Nankai trough in Japan ( Ando,1973; Shimazaki and Nakata,1980) . The same place boundary was ruptured twice subsequently, in two sets of delayed multiple events, the Ansei I and II events of 1854, and the Tonankai and Nankaido events of 1944 and 1946. In support of a time-predictable model of earth 5saka recurrence, Shimazaki and Nakata argued that the greater recurrence time between the first two sequences (147 years) and the second (91 years) is because the slip (and stress drop) were greater in 1707 than in either 1854 or 1946, the greater uplift at Muroto Point in 1707 (1.8 m) than in 1856 (1.2 m) or 1946 (1.15 m) being the evidence. The reason why this should happen is readily explained by the correlation between u and L.

Thus the ratio of fault length of the Hoei and Ansei II earthquakes, 500 km/300 km = 1.7 can explain the ratio of uplift at Muroco Point, 1.8/1.2 = 1.5 and recurrence cho, 147 /91 = 1. 6.

However, if this is interpreted as being due to a difference in dynamic stress drop, then one has to argue that a significant change in dynamic stress drop (50%) can occur on the same fault zone between succes-sive earthquakes. One could argue that this could occur because the slip in one earthquake might change the relative position of asperities on the

. 1 I

l 13 I stopping crack (Savage,1965 ). A stopping phase cannot physically stop the slip in these models because such a wave will lose energy with distance j whereas the results of the models are independent of dimension. A healing wave must be interpreted as a wave that propagates into the interior of the rupture in an analogous way, and f or analogous physical reasons , as the

' topping of cars on a highway propagates up the s trema of traffic.

s Causality reltricts it to travel at a velocity slower than a stopping phase. Thus Madariaga (1976, p. 648 ) observed, "I t appears as if a

' healing' wave propagates inward from the edge of the f ault some time af ter the P and S stopping phases."

~

Since slip is terminated by the healing wave, the rise time and final slip at any point on the fault is determined by the distance to the nearest boundary (Day,1979 ; Das ,1981 ). Therefore it is easy to see why mean slip ,

on a rectangular fault should be control'ed by the fault width.

A healing wave is the result df the boundary condir. ion that u = 0 at '

the edges of the fault. If the models are poor representations of large earthquakes, the most likely problem is that these boundary conditions are unrealistic. The models are of rectangular faults embedded ira an elastic whole space. The boundary condition u = 0 is imposed on all edges of the fault and healing waves -hus propagate from each edge. Since large earth-quakes rupture the f ree surf ace, slip is unconstrained there and a healing wave will not propagate from that edge. However even if an elastic half-space model were available, we would still expect slip to be width-dependent since it would be controlled by the healing wave from the base of the[ault.

In large earthquakes the base of the f aul t is at the bottom of the seismogenic layer. A plausible explanation for the seismogenic depth is .

14 that it is the result of a brittle-ductile trans ition. Thus a large l earthquake cannot propagate to greater depth because the energy at the crack cip is dissipated in plastic deformation. A more realistic model then may be one in which the base of the fault is in a plastic, rather than elastic, region and therefore the condition u = 0 is no. longer valid at that edge.

We illustrate in Figure S the difference between an elastic model and an elastic plastic model. The most significant difference is that in the elastic-plastic model (Figure 8b) slip at the base of the fault may be alloced to be greater than zero as a result of plastic deformation in a zone surrounding the rupture tip. This is simply the equivalent, in shear, of the blunting of a crack cip that occurs in tensile crack propagation in ductile materials. The plastic deformation around the base of the fault smooths out the stress singularity associated with finite slip there, and will continue as long as 'stip continues. This* may have the effect of inhibiting a healing wave from originating at the base, and if healing waves propagate only from the ends of the f ault, slip ar.d rise time will depend on fault length, not width.

No model is available with these boundary conditions but we can approximate one. If we make the approximation that slip stops abruptly with the arrival of the healing wave, then the final slip on the fault will be, from (7),

2 2 2 1/2 u(x,y) = (t h

~

(x 72 )) (9) v which we can calculate. This is a 'quasidynamic' model (Boacuright, 1980),

i.e. , a kinematic model that simulates a dynamic model.

7 l

15 It can readily be hown for the circular case that (9) yields final slip values that are evetywhere within 5% of that of the dynamic numerical l

models of Madariaga (1976) and Das (1980), and Day (1979) has shown that l

(9), when properly truncated, also yields a very good approximation to fl.nl slip in his rectangular models. We use it to simulate an elastic-plastic half-space model by simply assuming that no healing wave propa-gates from either the top or boccom of the fault.

The procedure we use is very similar to that used by Day (1979, pp. 23-26), and simply involves the calculation of th* "* ** * '" *d v = 0.98, for which the corresponding value of K is 0.81 (Dahlen, 1974),

and that the velocity of the healing wave is [IS. In Figure 9 ve show slip at the surface as a function of distance from the center of the fault for a bilateral case with L/W = 4. The mean slip is found to scale as i

_ gaod

, u= gL (10) so this model would lead to the interpretation that the linear correlacion between u and L that 'is observed means that the dynamic stress drop for large interplace earthquakes is appecximately constant. Equating (10) with (4) we obtain aa : 12 bars and 7.5 bars for thrust and strike-slip d

earthquakes, respectively. Returning to Figure 4, the line drawn through the data is the prediction of this model for ao = 10 bars. Furthermore, d

in this model, where slip is unconstrained at top and bottom, static stress drop will also be a function of fault length, since the scale length that determines the strain change will be the f aul t length. The observation made earlier that ao is a function of aspect ratio is due to the incorrect use of equation (2) to calculate it. According to this model, ao is also approximately constant for these earthquakes.

, - - - . , - - . 7

-.- , . - ---,-.e mnv-- .. - - , ,

t e

.0

> DISCUSSION The obs ervation that slip increases with fault length in large earthquakes poses severe consequences when viewed in the light of dynamic rupture models. In conventional dynamic models (W models), slip is deter-mined by fault width, rather than length. These modeia can only be recon-ciled with the observations if it is asstased that the dynamic stress drop determines the fault length, and the several major objections to this possibility were detailed earlier. With different assumptions concerning the boundary conditions at the base of the fault, it may be possible to construct a dynamic model in which slip depends on f ault length (L model).

This model avoids the objections raised to the W model but is based on a s peculativ,e , although not entirely ad hoc, assumption concerning the boundary conditions.

Furthermore, severe constraints are placed on L models from the geo-detic data obtained for the 1906 San Francisco earthquake. The simplest form of L model is one in which slip is totally unconstrained ac the base of the f ault. If this were the case, strain release would extend out to distances comparable to fault length, rather than depth, but as Brune (1974) has pointed out, the strain release in 1906 was concentrated within a few cans of las f rom th'e fault. From angle changes in the Pt. Arena triangulation network [ angle t from Thatcher (1975, Fig. 4)} one can

~$

estimate a strain drop of 8 x 10 t;ithin 12 km of the fault, a figure somewhat more consistant with a W model than an L model. Thus if L models are relevant, they must be models in which slip is only partially con-I strained at the base of the fault. In the absence of numerical modeling of

17 chis type, one can' e cell if this type of model will result in L scaling or  !

hybrid scaling intermediate to the L and W extremes.

These L and W models represent, in many respects, opposite extremes concerning the mechanism of large earthquakes and so it is us eful to discuss the contrasting way in which they scale. For earthquakes in which L < 2W, the models are indistinguishable in their gross manif es tations .

In Figure 10 we schematically show a comparison between an earthquake of  ;

dimensions about L = 2W and one of the' s ame width but about 15 ciaes longer. Specifically, this might be a comparison of the 1966 Parkfield earthquake, say, and the 1906 San Francisco earthquake. ,

On the lef t of the figure wu show a snapshot of slip on the fault during the smaller earthquake. We only show the part that is actually slipping during the snapshot. We also show the time history of slip at pome representative point. For simplicity, it is simply 'shown as a ramp with a rise time, t On the right is shown the predictions .of the two R.

models for the longer earthquake. -

In a bilateral case, as shown, the W model predicts that the slipping portion of the fault splits into two patches of length % W that propagate i away from each other at a velocity 2v as they sweep over the f ault surface.

Since the rise time t t W/28, remains the same but the slip is fifteen R

I times greater, the dynamic stress drop, and hence particle velocity, must be fifteen times greater.

i In the L model, the rupture sweepe out over the f ault as an expanding  !

patch, with slip continuing within its boundaries until after the finct dimensions are reached. In that model, the dynamic stress drop and par-ticle velocities are the same as in the smaller event, but the rise cias, t ge L/2 8 is much longer.

~

18 In terms of predicting the strong ground motions for a 1906 size earthquake, say, frem observed ground motions for a 1966 size earthquake, the difference between the W and L model is critical. The W model would predict that the average particle velocities would be much higher and the duration would be about the same. The L sodel would predict nearly the opposite.

Suppose we start with a square rupture of width W, and consider how peak particle 'relocity, u, p

and the asymptotic particle velocity, u,,

increase for ruptures of greater length. For a square rupture with dynamic S

stress drop, aed , the maximum value ofp u and the asymptotic value u, vill be

.S S -

0#

u = d o P

/ and (11)

.S S u

• 0# d o

Using the W model, for a rupture of width W, and length L > W , , the stress drop will have to be greater by the ratio de d ,

L 8 w o ac d

so that

( h * . Aa d 'W k dio p C l

19 (12)

W gg 8L u, e d W, For the L sodel, stress drop is the same but the scale length that deter-mines the maximum peak velocity becomes L rather than W, so that

,aad '

P and (13)

.L s 0#

u, = d Comparing (12) and (13), the two models differ in the ratios

• L , y D . _.0

W L P

and (14)

• L "o W

• W L "o

So that with a W model, from (12), both peak and asymptotic velocities for '

a 1906 type earthquake woull be about 15 times greater than for the Park-field earthquake. For the L model, from (13), the peak velecities would at maximum be about 4T grescer for a 1906 than a 1966 event, but the l

asymptotic value would be the sane.

l l

I

L 1

20 j l  !

These remarks, of course, apply only to the simple case of a smoothly propagating rupture. Any hacerogeneity will produce local high frequency variations in the velocities. However, they serve to point out the impor-tance of determining' if larr,e earthquakes are better described by an L -

model or W model or by scr.e intermediate case, if such can exist.

8 21 i

ACENOW EDCE!ENIS My attempts at trying to understand the consequences of slip corre-lating with fault length had a rather long gescation period, during which the author benefitted from discussions with T. Hanks, J. Boacwright, P.

Richards, S. Das, S. Day, and R. Madariaga.

Most of the work was done while the author was a visitor at the Department of Earth Sciences, Uni-versity of Cambridge, and a Green Scholar at the Institute of Geophysics and Planetary Physics, University of California, San Diego. Both are thanked for their support and hospitality. The work was supported by National Science Foundation grant EAR 80-07426 2nd Nacional Aeronautics and Space Administration grant NCR 33-008-146. I thank P. Richards and L.

Sykes for critical reviews.

Lamont-Doherty Geological Observatory con-tribution no. 0000. .

^

22 REFERENCES Abe, K. (1975). Reliable estimation of the seismic moment of large earthquakes, J. Phys. Earth, M, 381-390.

Aki, K. (1967) . Scaling law of seismic spectrum, J. Geoehys. Res ., 72,,

1217-1231.

Aki, K. (1972) . Earthquake mechanism, Tectonoehysics, 1_3,,

3 423-446.

Aki, K. (1980). Re-evaluation of stress drop and seismic energy using a new model of earthquake f aulting, in Source Mechanism and Earthquake Prediction, p. 23-50, Edit. Centre Nat. Recherche Sci., Paris.

Ambras eys , N. (1970). Some characteristics of the Anatolian faul t none, Tectonoehysics , 143-165.

Archuleta, R. J. (1976) . Experimental and numerical three dimensional simulations of strike-slip ear th quakes , Ph . D . thesis, U niv . of Calif. , San Diego.

Archuleta, R. J. , a nd S . M. Day (1980) . Dynamic rupture in a layered medium: the 1966 Parkfield earthquake, Bull. Seismol. Soc. Amer.,

20, 671-689.

Ando , M . (1975). Source mechanisms and tectcaic significance of his-coric earthquakes along the Nankai trough, Japan, Tectononhysics, M, 119-140.

Boatwright, J. (1980) . A spectral theory for circular seismic sources; simple estimates of source dimens ion, dynamic stress drop, and radiated seismic energy, Bull . Seismol . Soc. Amer., 70,, 1-27.

Bonilla, M. G. , and J. M. Buchanan (1970). Interim report on worldwide historic surf ace f aulting, U. S. Geol. Sury. Open-File Rept . , Was h-ington, D.C.

. \

1 23 Brune , J. N. (1974) . Current status s' understanding quasi-permanent fields associated with earthquakes, EOS, Trans. ACU, M, 820-827.

DaF,len, F. A. (1974) . On the ratio of P-wave to S-wave corner frequen-cies for shallow earthquake sources, Bull. Seismol. Soc. Amer. , M, a 59-1180.

Das, S. (1980). A numerical method f or determination of source-time functions for general three-dimensional rupture propagicion, Caoehys.

J. R. Astron. Soc., 62, 2 591-604.

Das , S. (1979) . Three-dimensional spontaneous rupture propagation and implications for the earthquake source mechanism, Geoch ys . J. R.

Astron. Soc., in press.

Day, S. (1979) . Three-dimensional finite difference simulation of fault dynamics , Final Rept., NAS2-10459, 71 pp., Systems, Science and Sof tware, La Jolla, Calif. .

Celler, R'. J. (1976). Scaling relations for earthquake source para-meters and magnitudes, Bull. Seismol. Soc. Amer., M, 1!01-1523.

Hanks, T. C. (1977) . Earthquake stress drops , ambient cectonic stress, .

and stresses elkat drive place motions, pure Appl. Geophys., M, 441-458.

Kanamori, H., and D. L. Anders on (1975). Theoretical basis of some empirical laws of seismology, Bull. Seismol. Soc. Am., M, 1073-1096.

Kostrov, 3. V. (1964). Selfsimilar problems of propagation of shear cracks , J. Appl. Math. Mech., g , 1077-1087.

Madcriaga , R. (1976). Dynamics of an expanding circular f ault, Bull.

Seismol. Soc. Amer., 6,6,, 639-666.  ;

Savage , J. C. (1965) . The stopping phase on seismograms , Bull . Seismol . j Soc. Amer. , g, 47-58.

.c--

24 Shimazaki, K., and T. Nakata (1980) . Time-predictable recurrence model l l

l for large earthquakes, Ceophys. Res. Lect., 7, 279-282.

Sieh, K. (1978) . Slip along the San Andreas fault associated with the great 1857 earthquake, Bull. Seis. Soe'. A=er., 68,, 1421-1448.

S lammons , D. 5. (1977). Scace of the arc for assessing earthquaka hazards in the United Scaces, Faults and earthquake magnitudes, U.S.

Army Eng. Waterway Exp. S ea. , Vicksburg, Mt.ss . , pp. 229, 1977.

Sykes, L. R., and R. C. Quictmeyer (1981). Repeac ciass of great ear thquake s along simple place boundaries, Third Maurice Ewing Symposium on Earchouake Prediccion, 4, edited by D. W. Simpson and P.

G. Richards , AGU, Washington, D.C.

Thatcher, W. (1975) . Strain accumulacion and release mechanism of the 1906 San Francisco earthquake, J. Geoohvs. Res., 80, 4862-4872.

Thatcher, W., and T. Hanka (1973) . Source parameters of southern Cali .

• fornia earthquakes , J. Ceophys . Res . , H, 8547-8576.

TAOLE 1 PAndHETERS Of LAHCE INTEMPLATE EAHTHQUAKES (AVEHACED FH0H SYKES Ato QUITTHEYEH (1981))

Ho L W u Ao No. Date Location 1027 d3ne-cm km km L/W om bars Strike-Slip Earthquakes

1. 10 3ul 1958 SE Alaska 4.3 350 12 29 '5 32 26

2. 9 3an 1857 S. California 7 380 12 32 465 36

3. 18 Apr 1906 San francisco 4 450 10 45 450 44

4. 19 May 1940 Imperial Va., Ca.

0.23 60 10 6 125 13

5. 27 Jun 1966 Parkfleid, Calif. 0.03 37 10 4 30 4

6. 9 Apr 1968 Ilorrego Htn, Ca. -

0.08 37 12 3 25 3

7. 15 Oct 1979 Imperial Va., Ca. 0.03 30 10 3 30 4

8. 4 feb 1976 Cuatemala 2.6 270 15 18 150 9

9. 16 Oct 1974 Cibbs F. l. 0.45 75 12 6 170 14

10. 26 Dec 1939 Ercincan, Turkey 4.5 350 15 23 285 18

11. 20 Dec 1942 Erbea liiksar, Turkey 0.35 70 15 5 112 8

12. I feb 1944 Cerede-ilolu, Turkey 2.4 190 15 13 275 16

13. 18 Har 1953 C5nen-Yenice, Turkey 0.73 58 15 4 280 21

14. 22 Jul 1967 Hudurnu, Turkey 0.36 80 19 5 100 7 Thrust Earthquakes

15. 6 11ov 1958 Etororu, Kurtles .

44 150 70 2.1 840 37

16. 13 Oct 1963 Eruppu, Kurtles 67 275 110 2.5 445 12

17. 16 May 1968 Tokacht-okl, Japan 28 150 105 1.4 355 10 18 11 Aug 1969 Shikotan, Kurtles 22 2 30 105 2.2 18 0 5

19. 17 Jun 1973 llemuro-oki, Japan 6.7 90 105 0.86 140 5

20. 4 llov 1952 Kamchat ka 350 45') 175 2.6 890 14

21. 28 Mar 1964 Prince Wm Sound, Alaska 820 750 180 4.2 1215 18 bl

22. 4 feb 1965 Itat Island, Aleutians 125 650 80 8.1 480 10 23.10 3an 1973 Colima, Hexico 3 85 65 1.3 110 5

24. 29 llov 1978 Daxaco, Mexico 3 80 70 1.1 110 5

25. 22 May 1960 S. Chile 2000 1(XX) 210 4.8 1900 21

26. 17 Oct 1966 C. Peru 20 80 140 0.6 360 12

i 26 FIGURE CAPTICNS Figure 1. Plot of log LW 2 V8- 108 M, for the large intraplace earthquakas from the data set of Sykes and Quittmeyer (1981). The lines of slope 1 are constant stress drop lines , aseuming c = 0.6 for the thrust events , and 0.3 for the strike-slip events.

Figure 2. A plot of mean slip, u, vs. f aul t length f or the strike-slip

~

events. The line drawn through the data has a slope of 1.25 x 10 '.

Numbers are references to Table 1.

Figure 3. The same as Figure 2, for the thrust events. The slope of the line is 2 x 10-5, Figure 4. A plot of log L W vs . log M,. The ic ne drawn through the data has a t.cpe of 1, for reference.

Figure 5. S tres s drop plotted vs. as pe ct ratio for the strike-slip earthquakes .

Figure 6. Stress drop vs. as pect ratio for the thrust ear thquakes .

Event 22 is an oblique slip event f;.: .'hich stress drop was calculated based only on the dip slip compnent and is hence underestimated.

Event 15 is an ancmalously deep even t in the Kuriles (Sykes and Qui ttmeyer , 1981) .

~

27 i

Figure 7. Schematic representation of two models of large earthquakes.

In A, it is represented by rupture in an elascic half-space. The boundary condicion ac the base of the rupeure is u = 0. In B, the rupture penetraces a duccile re gion. At the base u > 0, which is accommodated by plastic deformation in a zone surrounding the rupture tip.

Figure 8. Surface slip as a function of distance along the f aut e plane for two representative strike-slip earthquakes of similar width but different depch. Data for che Mudurnu earchquake is from Ambraseys (1969) and for the Ft. Tejon earthquake from Sieh (1978).

Figure 9. Dimensionless slip, u' vs . length, L', ac che free surface from the center to the end of the fault. The model is a quasidynamic .

one that simulates a dynamic model wich boundary condicions similar to those shown in Figure 7b, as described in the text. The normal-aa d

ization relacions are u = 7 Wu' and L = '4. ' . The case sMwn is bilateral wich aspect ratio 4.

Figure 10. A schematic diagram to illustrace che contrascing way in which a model in which width determines the slip (W model) scales with length as compared to a length dependent model (L model}.

i l

l l

l 1

l l

l

l - .

l I

23 I I I J g O

S THRUST 22 _

A STRIKE-SLIP [ q o *e 'eo/

00 7 6

21 _

• e E 20 - / / -

1 . / y

$ ,Q b

/ /

ig _ -

M

/

/$

Al l/

/

/ /,

18 - /p6 7

/

/

/

/

17 I i

~

26 27 28 29 30 Log Ma, dyne-cm Figure 1

0 0

C5 _ _ - 0 1

2 0

1 l

0 2 8 0

2 2

0 0

l 0

6 3

m e r

k 0 ,

u g

2 L i

. F 0

l 0

4 T 6 1

S 0 U 8 1

R 0 0 H l 0

T 7 2

6 1 5 94 1

1 2

6 3 2 2

- - /

0 0 0 0 0 0 0 0 5 0 2 1 1 ES"

500  ;  ; j -

' STRIKE- SLIP A2 A3 400 -

AI 300 -

A -

13 A 4 12 10 E

m 200 -

A A8 10 0 -

/15

/ ,6,7 I I I I O 10 0 200 _

300 400 500 L, km I

Figure 2

l . .

I i i I I 23 -

0 -

O O 22 g -

m E O

[p

~

9

, [g 21 -

g -

v f .

20 -

A A

19 - A A -

P 18 I I I I I 26 27 28 29 30 31 Log Ma , dyne-cm Figure 4

l . .

O l l l l C 4

_ o 7

- 4 _ o m

4 m s =

,J N 4 ^

_ o N

4 4

o_

t 4 9 #

<4 I I I I o

y o o o o to N -

SJ0q hp

_ _ _ _ _ _ _ _ - _ _ - - _ _ _ _ _ _ _ _ _ . _ _ _ _ _ _ _ m- _

40 i

~ ~

15 0 30

.8 20 - 8 '

E

• O g a 10 O O-22 CD 0 l

1 0 2 4 6 8 L/W Figure 6 l

k L

A

/

'7 BRITTLE

. _ _ - - - - -- - ~ ~ ~

'N DUCTILE NZONE OF PLASTIC DEFORMATION B

Figure 7

,, . . . _ _ - - , . . --,--c..-- - - --,--

- _ , - a g-MUDURNU,1967 3-a 10 -

O 50km FT. TEJON,1857

i a

_5 -

~

l l 1 0 10 0 200 300 km Figure 8 e

-..4 h

e 9

d r

I-

~

~h LLJ E J

l l

I i

l r0 N -

O l di 1S

. t

=

I t

a R

\ $

._ U U ;<

7 ,

L E L y E D

. O D 0

% O

/ 1 M M e

( -

r u

g

- i W .

L F

/ .

- 3n I

m

' /

t

=

I

/ "

l' .

u' nv '

x uQ

83G.9AP,HICAL, A SKETCH (MIOvifH[ THC *O* 5 SNG !NIOfruAleON FOn At L P(CFCSSION AL PCRONNEL e

68sOACEO ON THE PMOJECT. OEGINNING WITH THE PMINCIPAL INVE4flCAf 0R )

.NAME BIRTHDATE iuO. oAv. vai JAMES ti. BRUtlE November 23, 1934 PLACE OF BIRTH PRESENT NATIONALITY (C4TY, STATE. CObNT AY) (ALIENS BNOiCATE 44NO QF vasa ANO EXAAAfdON CATE)

, Modesto, California U.S. Citizen EDUCATION teEc4N wirN sacCALAUAEATE TRAININO AND INCLUDE PCSTOOCTCAAU DE*REE YEAR CONFEnRED INSTITUTION AND LOCATION 8,5c. 1956 University of flevada, Reno, Nevada Ph.D. 1961 Columbia University, flew York City HONOR 3 AND AWAADS tellow, GeoicgiTaf Society of America,1975.

President, Seismological Society of America,1970.

Grove Karl Gilbert Award in Seismic Geolo9y, ic67. .

Non.inated, tiew York Academy of Science,1966; Member,1970.

Fellcw, American Geopnysical Unicn, 1967.

First recipient of J.B. Macelwane Award by AGU,1962 MAJOR RESEARCV INTEREST Earthquake Source Mechanism Tectonics Earth Structure TN PaESENT POSITION. (fST PROFES$10NAL EACKGACUNO RESEARCH AND/OR PROFESSIONAL EXPERIENCE t,$,taatiNo Professnr of Geophysics - University of California, San Diego,1969 -

Associate Director, institute of Geophysics and Planetary Physics, University of California, San Diego, 1970-197C.

Chairman, Geological Research Division, Scripps Institution of Oceanography, University of

~

California, San Diego, 1974-1976. -

Associate Professor of Geophysics, California Institute of Technology, 1965-1969.

Adjunct Associate Professor of Geology, Columbia Unive'rsity,1964 Geop'ysicist, U.S. Ceast and Geodetic Survey,1964.

Research Scientist. Columbia University, 1958-1963.

Exploration Research, c.hevron Oil Company,1957.

Exploration Geophysics, Chevron Oil Company, 1956.

i 85 l 1

a urs.m. -

T BIBLIOGRAPHY James N. Brune s

1. (With J. Oliver) The Scismic Noise of the Earth's Surface, Bull. Seism.

Sbc. Amer., 49: 4, 349-353 (1959).

2. (With J. E. Nafe and J. E. Oliver) A Simplified Method for the Analysis and Synthesis of Dispersed Wave Trains, Jbur. Geophys. Res., 65: 1, 287-304 (1960).

3. (With J. E. Nafe) Observations of Phase Velocity for Rayleigh Waves in the Period Range 100 to 400 Seconds, Bull. Seiam. Sbc. Amer., 50: 3, 427-439 (1960).

4. Radiation Pattern of Rayleigh Waves from the Southeast Alaska Earthquake of 10 July 1958, Dcmin. Observ., 24,10, A Sympcsium on Earthquake Mechanism, 1-11 (1961).

5. (With M. Ewing and J. Kuo) Group and Phase Velocities for Rayle.igh Waves of Period Greater than 380 Secends, Sbience,133: 757 (1961).

6. (With J. E. Nafe and L. E. Alsop) The Polar Phase Shift of Surface Waves on a Sphere, Bull. Seism. Soc. Amer., 51. 247-257 (1961).

7. (With H. Benioff and M. Ewing) Long-pericd Surface Waves fEom the Chilean Earthquake of May 22, 1960, Racorded on Linear Strain Seismegraphs, Jbur. Geophys. Res., 66: 9, 2895-2910 (1961).

8. Attenuation of Dispersed Wave Trains, Bull. Seien. Sbc. Amer., 52:1, 109-112 (1962). .,

9. (With J. T. Kuo and M. Major) Rayleigh Wave Dispersion in the Pacific Ocean for the Period Range 20 to 140 Seconds, Bu'I. Seism. Sbc. Amer., 52:

2, 333-357 (1962).

10. Correcticn of Initial Phase Measurements for the Southeast Alaska Earthquake of July 10, 1958, and for Certain Nuclear Explosions, Jour. Geophys. Res.,

67: 9, 3643-3644 (1962).

I

11. (With M. Ewing and J. Kuo) Surface Wave Studies of the Pacific Crust and i Kantle, Geog. Moncgraph, 6, Crust of the R:cific Basin, (1962). l

12. (With J. Dorman) Seismic Waves and Earth Structure in the Canadian Shield, Bull. Seism. Sbc. Amer., 53: 1, 167-209 (1963).

13. (With A. Espinosa and J'. Oliver) Relative Excitation of Surface Waves by Earthquakes and Underground Explosions in the California-Nevada Region, l Jour. Ceophys. Res., 68: 11, 3501-3513 (1963).

14. Use of Surface Wave Rejection Filtors to Record Mantle Waves of Low Order, Earthquake Notes, 34: 73 (September - December 1963). (Abstract) 1

l l

James N. Bruns - Bibliography Page 2 .

1

~

15. (With P. W. Pomeroy) Surface Wave Radiation Patterns for Underground .

Nuclear Explosions and Small Magnitude Earthquakes, Jour. Geophys. Ros.,  !

68: 17, 5005-5028 (196?.).

16. Traval Times, Body Weyes, and Normal Modes cf the Earth, Bull. Seism. Soc.

Amer., 54: 6, 2099-2128 (1954).

i

17. (With A. Chander) Radiation Pattern of Mantle Rayleigh Waves and the Source Mechanism of the Hindu Kush Earthquake of July 6,1962, Bull. Seism.

Sbo. Amer. , 55 : 5, 805-819 (1965).

18. (With L. E. Alsop) Observations of Free Oscillations Excited by a Deep Earthquake , Jour. Geophys. Res., 70: 24, 6165-6174 (1965).

19. The Sa Phase from the Hindu Kush Earthquake of July 6,1962, Pure and Applied Physics, 62: 3, 81-95 (1965).

20. P and S Wave Travel Times 73d Spheroidal Normal Modes of a Homogeneous Sphere, Jour. Geophys. Ess., 71: 12, 2959-2965 (1966).

21. (With J. Oliver, A. Ryall and D. Slemmons) Micro-earthquake Activity Recorded by Portable Seismographs of High Sensitivity, BUZZ. Geol. Sbc.

of Amer., 56: 4, 899-Q24 (1966 ).

• 22. (With R. C. Liebermann, C. Y. King and P. W. Pomeroy) Excitation of Surface Waves by the Underground Nuclear Explosion Long Shot, Jour. Gacphys.

Res., 71: 18, 4333-4339 (1966).

23. (With C. R. Allen) A Micro-earthquake Survey of the San Andreas Fault System in Southern California, Bull. Seism. Soc. Amer. 57: 2, 277-296 (1967).

24. (With C. R. Allen) * *A Low-stress-drop, Low-magnitude Earthquake with Surface Faulting: The Imperial, Califernia, Earthquake of March 4, 1966, Bull. Seism.

Sbe. Amer., 57: 3, 501-514 (1967).

25. (With M. Wyss) The Alaska Earthquake of 28 March 1964: A Complex Multiple Rupture, Bull. Seism. Sbc. Amer., 57: 5, 1017-1023 (1967).

26. (With C. Y. King) Excitation of Mantle Rayleigh Waves of Period 100 Seconds as a Function of Magnitude, Bull. Seism. Sbc. Amer., 57: 6, 1355-1365 (1967).

26a. The Fault Slips, Engineering and Science Magazine, California Institute of Technology, 31: 2, 36-38 (1967).

27. Seismic Moment, Seismicity, and Rate of Slip along Major Fault Zones, Jbur.

Geophys. Res., 73: 2, 777-784 (1968).

P 1

James N. Brune - Bibliography Page 3

28. Source Dimensions of Earthquakes and Underground Explosions of Magnitude Near 4.0, Earthquake Notes, p. 22, (Abstract), June,1969.

29. (With C. R. Allen, A. Grants, M. M. Clark, R. V. Sharp, T. G. Theodore, E. W. Wolf and M. Wyss), The Berrego Mountain, California, Earthquake of April 9, 1968: A Preliminary Report., Bull. Seism. Soc. Amer., 58: 3, 1183-1186 (1968).

30. (With M. Wyss), Seismic Moment, Stress and Scurce Dimentions for Earthquakes in the California-Nevada Regien, Jour. Geophys. Res., 73: 14, 4681-4694 (1968).

31. Regional Variations in the Structure of the Upper Mantle and the Propagation of the Sa Phase, Centinental Margins and Island Arcs, Upper Mantle Comittee Symposi:er, Oc:.:.kz, Canada, IS65, GSC Paper 66-15, (1969).

32. Surface Waves and Crustal structure, Geophysical Monograph,13: 230-2u2 (1969).

33. (With G. R. Engen), Excitation of Mantle Love Waves and Definition of Mantle Wave Magnitude, Bull. Seism. Soc. Amer., 59: 2, 923-933 (1969).

33a. Seismicity, Rate of Slip, Stress and Heat Flow along the San Andreas f ault in Califctnia, EOS Trans. Amer. Geophys. Union, 50: 5, May 1969.

34 (With T. Henyey and R. Roy), Heat Flew, Stress and Rate of Slip Along the San Andreas Fault, California, Jour. Geophys. Res., 74: 15, 3821-3827 (1969).

35. (Wita W. Thatcher), Higher Mode Interference and Observed Anomaleus Apparent Love Wave Phase Velocities, Jour. Geophys. Res., 74: 27, 6603-6611 (1969).

36. (With *.. Trifur.ac)," Complexity of Energy Release During the Imperial Valley, Ca'.ifornia, Earthquake of 1940, Bull. Seism. Soc. Amer., 60: 1, 137-160 (1970).

37. (With D. Andersen, C. Archambeau, C. Richter, S. Smith), Earthquakes and Nuclear Detonations, Science,167: 1011-1012 (Feb. 13, 1970).

38. (With W. Arbas: and G. Engen), Locatiens of Small Earthquakes Near the Trifurcatien of the San Jacinto Fault Southeast of Anza, California, Bull.

Seism. Soc. Amer., 60: 2, 617-627 (1970).

' 39.

Tectonic Stress and the Spectra of Seismic Shear Waves from Earthquakes,

[ Jour. Geophys. Res., 75: 26, 4997-5009 (1970).

C. -

40. 5,elsmic Sources, Fault Plane Studies and Tectonics, SCS, 52: 5, 178-187, May 1971. (IUGG -Quadrennial Report on Seismology for U.S.)

f

. g ,(~

l' 'N - -

T

1 -

q ?\ p; V /Y {

r

James N. Brune - Bibliography Page 4

41. (with Wayne Thatcher) " Seismic Study of an Oceanic Ridge Earthquake Swarm in the Gulf of California'* Seaphys. J. R. astr. Soc. ,

22: 473-489 (July,1971).

-42. (with Cinna Lomnits, F. Mooser, C. R. Allen, and W. Thatcher) dSeismicity and tectonics of the northern Gulf of California Region, Mexico. Treliminary Results. " Seofisica Internacional,10:

37-48, Mexico, 1970.

43. " Seismic Methods for Monitoring Underground Nuclear Explosions, an Assessment of the Status and Outlook", (Book Review) Internatienal Institute fo'r Peace and Conflict Research (SIPRI) Stockholm, Sweden, Bull. Seism. Soc. Amer.

44.

(with W. Prothero, J. Dratler, B. Block) " Surface Wave Detecticn with a Broad-Band Accelerometer", ilature, 231:,21, 80-81 (May, 1971).

45. (with J. Davies) " Regional and Global Fault Slip Rates from Seismicity",

Nature, 229, 101-107 (January, 1971).

. 46. " Seismic Diser,imination Between Earthquakes and Underground Explcsions", ,

statement and testimony at Hearings before Subcommittee on Arms Centrol, International Law and Organization, Ninety-second Congress of the U.S. ,

• First Session on Comprehensive Nuclear Test Ban Treaty, 139-149 (July 22-23, 1971).

47. (with Max Wyss) " Regional Variations of Source Properties in Southern California Estimated from the Ratio of Short-to Long-Period Amplitudes",

Bull. Seism. Soc. Amer., 63,1153-1167 (October, 1971).

48. "A Deployment Program for Seismic Monitoring of a Comprehensive Test Ban Treaty", statement and testimony at Hearings before Subcommittee on Research, Development, and Radiation of the Joint Committee en Atomic Energy Congress of the U.S. , Ninety-Second congress, First Session on Extent of Present Capabilities for Detecting and Determining Nature of Underground Events, 133-142 (October 27-28, 1971).

49. (with W. Prothero) "A Suitcase Seismic Recording System", Full. Seism.

Soc. Amer., 6Z, 6,1849-1852 (December,1971).

50. (with D. McKenzie) " Melting on Fault Planes During Large Earthquakes",

Geophys. J.R. astr. Soc, 29:P (1972).

51. (with D. Oldenburg) " Ridge Transform Fault Spreading Pattern in Freezing Wax, Science, Vol. 178 (1972) 301.

• James N. Brune - Bib 11egraphy Pago 5 l 52. C. R. Allen, M. Wyss , J. N. Brune, A. Grants and R. E. Wallace.

" Displacements on the Imperial, Superstition Hills , and San Andreas Faults Triggered by the Borrego Mountain Earthquake". In U.S.G.S., Prof. Paper

$787, pp.87-104 (L972 ).

l

[ 53. B. E. Tucker and J. N. Brune. " Seismograms, S-Wave Spectra and Source l Parameters for Aftershocks of the San Fernando Earthquake of February 9, 1971." NCAA SpeciaL R.tporC ,1973.

54. I. Reid, M. Reichle , J. Brune and H. Bradner. "Microcarthquake Studies using Sonobuoys: Preliminary Results from the Gulf of California."

Geophys. J. R. aser. Soc., E, 365-379 (1973).

2

55. J. N. Bmne , S. de 1,a Cruz, H. Bradner, C. Villegas , I. Reid, M. Reichle, A. Nava, M. Lozada and P. Silva. " Earthquakes in the Gulf of California Recorded using Land-Based Recordings of Moored Hydrophone Arrays."

Geofisica Inc. , E (3), 201-212 (L972 ) .

, 16. J. N. Brune and C. Lomditz. "Recent Seismological Developments Relating

, to Earthquake Hazard." Geofisica In% 14: pp. 49-63 (1974).

57. P. Molnar, B. E. Tucker and J. N. Brune. " Corner -Frequencies of P and S Waves

&~ Models 'of Earthqu'aRe Sources,"' Bull. Seismo. Soc. Am., 63, 2091-2105 (1973).

58. F. Gilbert, A. Dziewonski and J. Brune. d An Informative Solution to a

• Seismological Inverse Problem". Proc. #ct 'I. Acad. Sci., E, 5, pp. 1410 ( 197't .) .

59. W. Thatcher and J. N. Brune. " Surface waves and crustal structure in thet Gulf of' California region." Sull. Seism. Soc. Am, 63_, 5, 1689-1693*(1973).

60. Brune, J. N. . " Earthquake modelling by stick; slip along pre-cut surfaces in stressed foam rubber". Sull. Seism. Soc. Am. , 63, -No. 6. , 2105-2119.

(1973). .

61. Brune , J. N. and F. Gilbert, " Torsional Overtone Dispersion from Correla-tions of S Waves to SS Waves", Bull. Seism. Soc. Am., 64 (2), 313-320

-(1974).

62. H. Bradner and J. Brune, "The Use of Sonocuoys in Determining Hypocenters of Aftershocks of the February 21,.1973 Pt. Mugu Earthquake ," Sull. Seism. Soc.

. Am., 6_4,, No. 1,99-101, 1974.

63. J. N. Brune, " Current Status of Understanding Quasi-Permanent Fields Associated with Earthquakes" EOS, 55, No. 9, 1974.,

64. D. W. Oldenburg and J. N. Brune, "An Explanation for the Orthogonality of Ocean Ridges and Transform Faults", J. Geophys. Res., 80,, no.17,

..p. 2575, 1975.

65. Alfonso Reyec, J. Brune, L. Canales, J. Madrid, J. Rebo11ar, L. Munguia, T. Barker, "A Microcarthquake Survey of the San Miguci Fault,. Baja California, Mexico',' Geophys. Res. Lttrc. , 2_, 56-59, 1975.* -

66. Brune, J. N., The Physics of Earthquake Strong Motion, in Seismic Risk and Engineering Decisions. C. Lomnits and E. Rosenb1ueth (eds. ), Elsevier Sci.

Publ. Co., New York. 1976.

e r - e - s-,-

~

l

- James N. Bruna - Bib 11ogrrphy P go 6 l l

• i l

664. James Brune, Cinna Lomnits, Clarence Allen, Frederico Mooser, Francis Lehner, l and Alfonso Reyes,"A Permanent Seismograph Array Around the Gulf of l California " ru m sein,. Ma. Am. , 66, 969-978, 1976.

67. Ralph Archuleta and James N. Brune, " Surface Strong Motion Associated with a Stick-Slip Event in a roam Rubber Model of Earthquakes," sun ,

Seismo. Soc. Am. , j5,,1054-1971, 1975. ,

, 68. Brian E.' Tucker and J. N. Brune, " Source Mechanism and Surface-wave

, Excitation for Aftershocks of the San Fernando Earthquake", Geophys.

J. R. astr. Soc., u , 371e426,.1977, -

s 69. Michael Reichle,, George Sharman, and James Brune,"Sonobuoy and Teleseismic Study of Two Gulf of California Transform Fault Earthquake Sequences",

,. Sull. Seismo. Soc. Amer., 6,6,, 1623-1642, 1976.

70. ' William A. Prothero, Ian Reid, Michael Reichle , James Brune ,10cean -Bcttom Seismic Measurements on the East Pacific Pdse and Rivera Fracture Zone",

Nature, 262, 121-124, 1976.

' 71. George F. Sharman, Michael . S. Reichle, James. N. Brune , "A Detailed Study of Relative Plate Motion in the Gulf of California" Geology, April, pp. _

206-210, 1976. ,

72. Stephen H. Hartzell and James N. Brune, "Scurce Parameters for the January, 1975 Brawley - Imperial Valley Earthquake Swarm" PAGEOPH, L:.5 p, 333 - 355, 1977.

73. James N. Brune, Alfonso Reyes ,' Michael S. Reichle, "Recent Seismic and Tectonic Studies of the Gulf of California", CIBCASIO Annual . deport, 1976. .

74. James N. Brune, R. Archuleta, J. Frazier, G. Hegemier, " Physical and Numerical Modeling of Spontaneous Slip", su - y of talk given at Northwestern University at NSF Workshop on " Application of Elastic -

Waves in Electrical Devices, Non-Destructive Testing and Seis=61ogy" May 24-26, 1976. ,

75. James N. Brune,"Q of Shear Waves Estimated from S - SS Spectral Ratios," Geophys. Res. Lttrs. , 4, No. 5,1977.

'.76 Stephen.H. Hartsell, Gerald A. Frazier and James N. Brune, " Earthquake modeling in a homogeneous half space," Bull. Seism. Soc. Am. , 68, 301-316, 1978.

77. Keith Priestley and James N. Brune, " Surface Waves and the Structure of the Great Basin of Nevada and Western Utah", J. Gaophys. Res. , 83 2265-2272,'1978.

, 78. Luis Hunguia, M. Reichle, A. Reyes, R. Simons, J. F. Brune. "Aftershocks of the 8 July u /5 Canal De Las Ballenas, Gulf of California, Earthquake",

Geuphysical Res. Ltte ., 9_, No. 11, 1977.

g -

e

d.79. J. N. Brune, " Implications of Earthquake Triggering and Rupture Prepa-gation for earthquake Prediction Based on Premonitory Phenomena",

presented at USGS Conference on Fault Mechanics and its Relation to Earthquake Prediction, December 1-3, 1977, J. Geophys. Res., 84, BS, 1979.

l

./ 80. J. N. Brune, R. J. Archuleta and S. H. Hart: ell, "Far-Field S-Wave l

Spectra, Corner Frequencies and Pulse Shapes", presented at USGS  ;

Conference on Fault Mechanics and its Relation to Earthquake Prediction i December 1-3, 1977, J. Geophys. Res., 84, B5, 1979. .; - .

81. Stephen Hartzell, James N. Brune and Jorge Prince, "The October 6, 1974 Acapulco Earthquake and the Importance of Short P:riod Surface Waves in Strong Ground Motion, Bg . sais o. M . Ag , 68,, 1663-1677, 1978.

82. James N. Brune, " Statement to the ACRS", meeting of the Subcommittee of the Advisory Committee on Reactor Safeguards, Los Angeles, California, June 23, 1977.

83. B. Tucker, J. King, K. Aki, J. Brune, I. Nersesov, W. Prothero, G. Shpilker and G. Self, "A Preliminary Report on a Study of the Seismic Respense of Three Sediment-Filled Valleys in the Garm Region 'of the USSR," Proc. Second Int'l Conf. on Micronization for Safer Construction Research and Application, pp. 10S1-1062, San Francisco, California, USA Nov. 26-Dec. 1 1978.

84 A. Reyes, J. N. Brune and C. Lemnits, " Source Mechanism and Ar,tershock

, Study,of the Colima, Mexico Earthquake of January 10, 1973", Bull.

Seism. Soc. of Amer., 69, 6, pp. 1819-1640, 1979;

85. Stephen Hartzell and James N. Brune, "The Horse Canyon Earthquake of .

August 2, 1975 - Two Stage Stress Release Process in a Strike-Slip Earthquake", Bull. Seism. Soc. Am. , 39, No. 4, 1161-1173, 1979.

86. J. N. Brune, " Cooperative Seismic Studies in Mexico", Invited paper, INTERCIENCIA/AAAS Symposium, Houston Texas, January 3-8, 1979.

87. J. N. Brune, " TESTIMONY" before the Nuclear Regulatory Commission, San Luis Obispo, California January 9, 1979

88. J. N. Brune, M. Hernandez, J. Gonzale=, L. Munguia, R. Simons, . Suare and F. Vernon, " Digital Seismic Event Recorders: Cescription and Exauples from the San Jacinto Fault, the Imperial Fault, the Cerro Prieto Fault and the Oaxaca, Mexico Subduction Fault", in press, 1980 A ,7 ( ,

89. Albores , A . A. Reyes , J. N. B' rune , J. Gonv.alez, L. Garcelazo, F. Suarez,

" Seismicity Studies in the Ferton of the Cerro Prieto Geothetunt Field".

Proceedings of the First Symposium on the Cerro Prieto Geothermal Field.

September 20-22, 1978, San Diego, California.

90. A. Leeds and i. N. Brune, "The Locations of the 1954 Northern Baja Californi c Earthr:unke,

91. L. Munguia, J. N. Brune, A. Reyes, J. Gonzalez, R. Simons, F. Vernon,

" Digital icismic Event Recorder Records and Spectra for Af tershocks of the November 29, 1978 Oaxaca Earthquake", Geofisica Internacional, 17, 3, 359-366,'(1978). -

I l

i

92. J. N. Brune, R. 5. Simons, C. Rebollar and A. Reyes, " Seismicity and Faulting in Nor".hern Baja California" in,1979 National Geological Society of America Fle'.d Trip Guidebook, Geological Hazards in the San Diego Area, Meeting held November 5-8, 1979, San Diego, California.

93. L. C. Pakist.r and J. N. Brune, " Roots of the Sierra Nevada", Science, 210, pp. 10'38-1094, (5 December b80).

9"* Keith Priestley, James N. Brune and John A. Orcutt, " Higher Mode Surface Waves and Structure of the Great 5asin of Nevada and Western Utah" ('J. Geophysics 1 Tesearch , 85, B12, 7166-7174, Decenber 10, 1980.

95. James Brune, Jorge Prince, Frank Vernon III, Enrique Mena, Richard Simons, Strong Motion Data Recorded in Mexico for the October 15, 1979 Imperial

, Valley Earthquake. In press, USGS, 1980.

96. F. Alejandro Nava and James N. Brune, An Earthquake-Explosion Reversed Refraction Line in the Peninsular Ranges of Southern California and Baja California Norte. In preparation, 1980. ,,,, w 3.,.,.,,

97. F. Alejandro Nava and James N. Brune, Source Mechanism and Surface Wave Excitation for Two Earthquakes in Northern Baja California, Mexico. In preparation, 1980. j, , . , , ,

98. Michael S. Reichle, Keith Priestley, James Brune and J. A. Orcutt, The 1973 Oaxaca Earthquake Source Mechanism Analysis frem Digital Data. Geofisica Internacional, 17, 3, 295-301, (1973).

99. Dave Chavez, Javier Gonzales, James N. Brune, Frank Vernon III, Richard Simons, L. K. Hutton, Peter T. German, Carl E. Johnson, Mainshock Location and Magnitude Determination Using Combined U.S. And Mexican Data. In preparation, USGS, 1980.

100 Reyes, A., J. Gonzales, L. Munguia, A. Nava, F. Vernon, J. N. Brune.

Locations of aftershocks of the Oaxaca earthquake using smoked paper recorders and digital event recorders. Geofisica Internacional,

]L[,

j 3, 341-358, (1978).

101. Testimony before ASLB Appeal Board re Diablo Canyon (8/8/80).

102. Testimony on behalf of Gov. Brown re ASLB Appeal Board Question 7 on Diablo Canyon (8/8/80).

103 King, Jerry and James N. Brune. Modelling the seismic response of sedimentary basins. Submitted to BSSA (1980).

104 Affadavit' for Atomic Safety and Licensing Board of NRC (2/29/80) l (filed in order to reopen Diablo Canyon hearings) 105. ACE Transcript of Testimony before Atomic Energy Commission (3/15/73)

(Direct and Cross Examination) ( ACE-Federal Reports, Inc./ Official Reporters /415 Second St., N.E./ Washington, D. C. 20002/(202 547-6222) l

106. Keith I'. Priestley and James N. Brune. Shear Wave Structure of the Southern volcanic Plateau of Oregon and Idaho, and Northern Great Basin of Nevada from Surface " dave Dispersion.

(In progress, 1981).

107. Luis Mendoza G., Alfonso Reyes Z., Michael S. Reichle, James N. Brune.

La Geofisica del Golfo de California: una revision. (In progress, 1981).

108. Luis Hunguia and James N. Brune. High Stress Drop Events in the Victoria, B.C. , Mexico, Earthquake Swarm of March,1978.

(In progress, 1981).

109. Affadavit before Atomic Safety and Licensing Board of NRC (5/22/81) l e .

e a

e

e . 1 LIST OF REFERENCES Aki, K., 1978, Origin of the Seismic Gap: What initiates and stops a rupture propagation along a plate boundary: Proceedings of Fifth National Meeting of the Universities Council for Earth-quake Engineering Research, June 23-24, 1978; Mass. Institute of Technology.

Albee, A. L. and Smith J. L., 1966, Earthquake characteristics and fault activity in southern California, in Lung, R. and Proctor, R., Eds. Engineering Geology in Southern California:

Association of Engineering Geologists, Glendale, CA., pp.

9-33 Allen, et al. 1967, Report to Stewart L. Udall, Secretary of the Interior, Geological and seismological factors pertaining to the proposed construction of a nuclear power-desalting plant at Bolsa Island, California, p. 18

---

, 1975, Geological criteria for evaluating seismicity:

Geological Society of America Bulletin, v. 86, p. 1041-1057 Ambraseys, N.1., 1978, Preliminary analysis of European strong-motion data 1965-1978 Part II: Proceedings of European Association of Earthquake Engineering, Fall, 1978 Meeting

. held in Dubrovnik, Yugoslavia, September, 1978.

~

Anderson, J.G., 1979, Estimating the seismicity from geological' structure for seism'ic risk studies, Seismological Society of America Bulletin, v. 69, p. 135-158 Angstman, B.G., Spudirh, P.F.K. and Fletcher, J., 1979, The Coyote Lake Earthquake: 0.42g acceleration from an S-P converted phase: EOS, Transactions, American Geophysical Union, v. 60, p. 890 Archuleta, R. J. and Brune, J. N., 1975, Surface strong motion associated with a stick-slip event in a foam rubber model of earthquakes: Seismological Society of America Bulletin,

v. 65, p. 1050-1071

---

and F ra zier , G . A . , 1978, Three-dimensional numerical sLmulations of dynamic faulting in a half-space: Seismological Society of America Bulletin, v. 68, p. 541-572.

---

, 1979, Rupture propagation effects in the Coyote Lake earthquake. EOS, Transactions, American Geophysical Union, v. 60, p. 890

---

and Spudich, P., 1981,Anexplanationfor[helarge amplitude vertical accelerations generated by the 1979 Imperial Valley, California earthquake: 1981 Spr-ng AGU Meeting, EOS, v. 62, p. 323.

(

Bakun, W. H., Stewart, R. M. and Bufe, C. G., 1978, Directivity in the high-frequency radiation of small earthquakes:

Seismological Society of America Bulletin, v. 68, p. 1253-1264.

(

Blandford, R. R., 1975, A source theory for complex earthquakes:

Saismological Society of America Bulletin, v. 65, p. 1385- ,

1405. '

Boatwright, J., 1980, Preliminary body-wave analysis of the St. Elias, Alaska earthquake of February 28, 1979:

Seismological Society of America Bulletin, v. 70, p. 419-436 Boore, D. M., 1977, Strong motion recordings of the California earthquake of April 18, 1906: Scismelogical Society of America Bulletin, v. 67, p. 561-578.

- - - - - - - - - - - , Joyner, W. B., Oliver, A.A., ard Page, R. A., 1978, Estimation of ground motion parameters: U. S. Geological Survey Circular 795, 43 p.

---

,Porcella, R. L., (1980), Peak acceleration from strong motion records: a postscript: Seismological Society of America Bulletin, v. 70, p. 2295-2297.

Brazee, R. J. and Cloud, W.K.,

1958, U.tited States Earthquakes, 1956: U. S. Department of Commerce, U. S. Government Printing Office, Washington, D.C. ,

Brune, J. N., Espinosa, A., and Oliver, J., 1963, Relative excitation of. surface waves by earthquakes and underground explosives in the California-Nevada region: Journal of Geophysical Research, v. 68, p. 3501-3513.

---

, Simons, R. S., Rebollar, C. and Reyes, A., 1979, Seismicity and faulting in northern Baja California, in Abbott, P. L. and Elliott, W. J., Eds., Earthquakes and Other Perils - San Diego Region: Geological Society of America Annual Meeting, San Diego, California, 1979, Field Trip Guidebook, p.83-100.

---

, Prince, J., Vernon, F. L., Mena, E. and Simons, R.S.,

1980, Strong motion data recorded in Mexico during the October 15 main shock: In Press, USGS Professional Paper on the Imperial Valley Earthquake of October 15, 1979.

Campbell, K. W., 1980, Attenuation of peak horizontal acceleration within the near source region of moderate to large earthquakes, TERA Technical Report 80-1, TERA Corporation, Berkeley, California.

l

- - __ . . -=

s  :. .

5 Crouch, J. K., 1979, Neogene tectonic evolution of the California continental borderland and western transverse ranges: Geological Society.of America Bulletin, v. 90, p. 338-345 Federal Emergency Management Agency (FEMA) From Analyses Carried

-Out by the National Security Council ad hoc Committee on assessment of consequences and preparations for a major California earthquake, 1981, "An Assessment of the consequences .

+

and Preparatio:a for a Catastrophic California Earthquake":

Findings and Actions Taken: M-& R 2, Washington, D. C.

1 Fletcher, J. B..Brady, A. G., and Hanks, T. C., , S trong motion accelerograms of the Oroville, California Af tershocks:

Data processing and the aftershock of 0350 August 6, 1975:

Seismological Society of America Bulletin, v. , p.

4 Greene, H. G. and Kennedy, M. P., 1980, Review of offshore seismic

' reflection profiles in the vicinity of the Cristianitos fault, San Onofre, California: U. S. Department of the Interior Administrative Report (Appendix F, Safety Evaluation Report, NUREG-0712)

, Hanks, T. C., 1974, The faulting mechanism of the San Fernando earthquake: Journal of Geophysical Research, v. 79,

p. 1215 - 1229.

____________and Thatcher, W., 1972, A graphical representation of seismic source parameters: Journal of Geophysical Research, v. 77, p. 4393-4405.

---

, Hileman, J. A., and Thatcher, W., 1975, Seismic moments of the larger earthquakes of the southern California region: Geological Society of America Bulletin, v. 86,

p. 1131-1134.

---

, and Johnson, D. C., 1976, Geophysical assessment of peak accelerations: Seismological Societylof America Bulletin,

v. 66,1p. 959-968 f

---------- , and McGuire, R. K., 1981, The character of high frequency strong ground motion

Preprint

- - - - - - - - - - , and Kanamori, H., 1979 , A moment-magnitude scale:

Journal of Geophysical Research, v. 84, p. 2348 - 2350.

Hartzell, S. H. and Brune, J. N., 1977, Source parameters for the January, 1975 Brawley-Imperial Valley earthquake swarm

Pageoph.

v. 115, p. 333-355.

t i

p j_ '

l

l l

l Hartzell, S. H., Brune, J. N., Prince, J., 1978, The 9ctober 6, 1974 Acalpulco earthquake: An example of the importance of short period surface waves in strong ground motion: Seismo-logical Society of America Bulletin, v. 68, p. 1663-1677.

- - - - - - - - - - - , and Helmberger, D. V., 1980, Strong ground motion modeling of the October 15, 1979 Imperial Valley earthquake:

Transactions, American Geophysical Union, EOS, 1980 Fall AGU Meeting, v. 61, p. 1036.

Heaton, T. H., 1978, Generalized ray models of strong ground motion,

! Ph.D. Thesis, California Institute of Technology, Pasadena, California.

---

, and Helmberger, D. V., 1979, Generalized ray models of the San Fernando earthquake: Seismological Society of America Bulletin, v. 69, p. 1311 - 1342.

House, L., and Boatwright, J., 1980, Investigation of two high stress drop earthquakes in the Shumagin seismic gap, Alaska:

Journal of Geophysical Research, v. 85, p. 7151-7165.

Joyner, W. B., Boore, D. M. and Procella, R. L., 1981, Peak horizontal acceleration and velocity from strong-motion records including records from the 1970 Imperial Valley, California earthquake: U. S. Geological Survey Open File Report 81-365, 46 p. - -

---

, Boore, D. M. and Porcella, R. L., 1981f Peak horizontal acceleration and velocity from strong-motion records: Seismological Society of America Abstracts, 76th Annual Meeting, v. 52, p. 80-81.

Kanamori, H. and Anderson, D. L., 1975, Theartical basis of some empirical relations in seismology: Seismological Society of America Bulletin, v. 65, p. 1073-1095.

- - - - - - - - - - - , and Regan, J., 1981, Long-period surface waves generated by the Imperial Valley Earthquake of 1979:

In Press, U. S. Geological Survey Professional Paper on Imperial Valley Earthquake.

Lachenbruch, A. H. and Sass, J. H., 1980, Heat flow and energetics of the San Andreas fault zone: Journal of Geophysical Research,

v. 85, p. 6185-6223.

Legg, M. R., and Kennedy, M. P., 1979, Faulting offshore San Diego, California and northern Baja California Mexico, in Abbott, P. L. and Elliott, W. J., Eds., Earthquakes and Other Perils -

San Diego Region: Geological Society of America Annual Meeting, San Diego, California, 1979, Field Trip Guidebook, p. 29-46.

I 4-

. ~

I l

Luco, J. E., 1980, Review of the report " Simulation of earthquake ground motions for San Onofre Nuclear Generating Station, Unit I. Supplement III, August, 1980." A report prepared for the Nuclear Regulatory Commission.

McGarr, A., 1981, Analysis of peak ground motion in terms of a model of inhomogeneous faulting: Journal of Geophysical Research, v. 86, p. 3901-3912.

Miller, R. K., and Felszeghy, S. F., 1978, Engineering features of the Santa Barbara earthquake: Earthquake Engineering Research Institute Report UCSB-ME-78-2.

Moore, G. W. and Kennedy, M. P., 1975, Quarternary faults at San Diego Bay, California: U. S. Geological Survey Journal of Research, v. 3, p. 589-595.

Mueller, C. S., and Boore, D. M., 1981, Site amplification at El Centro strong motion array station #6: Earthquake Notes,

v. 52, p. 84.

Nava, F. A., 1980, Source mechanism and surface wave excitation for two earthquakes in northern Baja California Mexico:

Ph.D. Thesis, University of California, San Diego'.

---

, and Brune, J. N., 1981, Source mechamism and surface wave excitation for two earthquakes in northern Baja California Mexico: Submitted to Geophysical Journal.

Panel on Earthquake Problems Related to Siting Critical Facilities, 1980, Earthquake Research for the Safer Siting of Critical Facilities: National Academy of Sciences, Washington, D. C.,

48 p.

Porter, L. D., Ragsdale, J. T., McJunkin, R. D., 1979, Processed data from the strong motion records of the Santa Barbara earthquake of 13 August 1978: Special Report 144, California Division of Mines and Geology, v. 3.

Reagor, B. G., Stover, C. W., Algermissen, S. T., Steinbruge, K. V.,

Hubiak, P., Hopper, M. G., and Barnhard, L. M., 1981, Preliminary evaluation of the distribution of intensity: U. S. Geological Survey Open File 80-1094 Sbar, M. C. and Sykes, L. R., 1973, Contemporary compressive stress and seismicity in eastern North America: An example of intra-plate tectonics: Geological Society of America Bulletin, v. 84,

p. 1861-1882.

i i

l

r

.- . e Scholz, C. H., 1981, Scaling ' laws for large earthquakes consequences for physical models: (Unpublished) Department of Geological Sciences and Lamont-Doherty Geological Observatory, Columbia University, Palisades, New York, 37 p.

Simons, R. S., 1977, Seismicity of San Diego, 1934-1974: Seismological Society of America Bulletin, v. 67, p. 809-826.

---

, 1979, Instrumental seismicity of the San Diego Area, 1934-1978, in,~ Abbott, P. L. and Elliott, W. J., Eds., Earth-quakes and Other Perils - San Diego Region: Geological Society of America Annual Meeting, San Diego, California, 1979, Field Trip Guidebook, p. 101-105.

---

, Brune, J. N., Anderson, J. G., 1981, The Guadalupe Victoria strong motion record from the near field of the June 9,1980 northern Baja California earthquake: Seismological Society of America Abstracts, 76th Annual Meeting, v.

~

52,

p. 85-86.

Slemmons, D., 1977, State-of-the-Art for assessing earthquake hazards in the United States; Report 6; Faults and Earth-quake Magnitude: Mackay School of Mines, University of Nevada, Reno, Nevada, U. S. Army Eng. Waterways Experiment Station, 129 p.

R. A., Cline, M. W., and Timmerman, E. L., 1981, Horizontal Sn'ay, deformation in the Imperial Valley between 1934 and 1980:

(Unpublished) Submitted to Journal of Geophysical Research.

Sykes, L. R. and Quittmeyer, 1981, Repeat times of great earth-quakes along simple plate boundaries. Third Maurice Ewing Symposium on Earthquake Prediction, 4, Jimpson, D. U. and Richards, P. G., Eds., American Geophysical Union, Washington, D. C.

Swanger, H. J., Day, S. M., Murphy, J. R., Guzman, R., 1978, State-of-the-art study concerning near-field earthquake ground motion: Systems, Science and Software, NUREG/CR-1978.

Thatcher, W. and Brune, J. N., 1971, Seismic study of an oceanic ridge earthquake swarm in the Gulf of California: Geophysical J. R. Astr. Society, v. 22, p. 473-489.

---

, 1972, Regional variations of seismic source para-meters in the northern Baja California area: Journal of Geophysical Research, v. 77, p. 1549-1565.

- Trifunac, M. D., and Brune, J. N., 1970, Complexity of* energy release during the Imperial valley, California, earthquake of 1940: Seismological Society of America Bulletin, v. 60, p . 13 7- 16 0.

t

Trifunac, M. D., 1972, Stress estimates for the San Fernando, California earthquake of February 9, 1971: Main Event and 13 aftershocks: Seismological Society of America Bulletin,

v. 62, p. 721-750.

- - - - - - - - - - - - , 1972, Tectonic stress and the source mechanism of the Imperial Valley, California, earthquake of 1940:

Seismological Society of America Bulletin, v. 62, p.

1283-1302. .

---

, and Brady, A. G., 1975a, On the correlation of seismic intensity scales with the peaks of recorded strong ground motion: Seismological Society of America Bulletin, v. 65, p. 139-162.

U. S. Geological Survey, 1981, Scenarios of Possible Earthquakes Affecting Major California Population Centers, with Estimates of Intensity and Ground Shaking: Open File Report 81-115, 36p.

Wentworth, C. M., Bonilla, M. G., and Buchanan, J. M., 1973, Seismic environment of the sodium pump test facility at Burro Flats, Ventura County, California: U. S. Geological .

Survey Open File Report 73-360.

Woodward & Clyde Consultants, 1978, Geotechnical evaluation of potential island and offshore California LNG Import Terminal Sites: Report to California Coastal Commission, Sacramento, California.

Wyss, M. and Brunce, J. N., 1971, Regional variations of source properties in southern California estimated from the ratio of snort- to long-period amplitudes: Seismological Society of America Bulletin, v. 61, p. 1153-1167.

. l l

1 l