ML20009A905
| ML20009A905 | |
| Person / Time | |
|---|---|
| Site: | San Onofre  |
| Issue date: | 06/22/1981 |
| From: | Atomic Safety and Licensing Board Panel |
| To: | |
| References | |
| ISSUANCES-OL, NUDOCS 8107140446 | |
| Download: ML20009A905 (65) | |
Text
mm wnm,
-~
QQtglN AL
^
NCC'-'M REGUI.ATORY COMMISSICN
~
Y t I ? L, /,.
- p s fs h
ATOMIC AND SAFETY LICENSING BOARD /
\\
y H._
9 4
O i ?-
f/
$/
/ /
In the Matter cf:
)
SOUTHERN CALIFORNIA EDISON COMPANY,
)
DOCKET NOS. 50-361, OL ET AL.,
)
and 50-23E, OL (SAN ONOFRE NUCLEAR GENERATING
)
762.f STATION, UNITS 2 AND 3
)
CATE:
June 22, 1981, ff.
pAggg:
A;;
San Diego, California 7gol 5
APPLICANTS' DIRECT TESTIMONY l/o Oi.55 7
b.
W - 10 ON L
ppd-fo g
s
!O O
E'F CONTENTION NO. 1
- 5. Hlkot"gy '.
Y'), e W
m O
A, M' O
/^
.uneasox arrowrun o
M 400 Virgini.1 A*re., S.*d.
Wasni..g On, C.
C.
200 4 f
Te ' a=hene : (202) 554-2245
=
0107140446 810622 PDR ADOCK 05000361 T
PDR e
O I
UNITED STATES OF AMERICA
( -)
2 NUCLEAR REGULATO'RY COMMISSION O
3 BEFORE THE ATOMIC SAFETY AND LICENSING BOARD 4
5 In the Matter of
)
Docket Nos. 50-361 OL
)
50-362 OL 6
SOUTHERN CALIFORNIA
)
EDISON COMPANY, E @.
7 (San Onofre Nuclear Generating
)
Station, Units 2 and 3)
)
O 9
10 11 12 13 APPLICANTS' DIRECT TESTIMONY OF DR. STEWART W.
- SMITH, 14 DR. GERALD A.
FRAZIER, DR. SHAWN BIEHLER, DR.
I.M.
IDRISS AND DR. ROBERT L.
MCNEILL 15 ON CONTENTION #1:
16 O
18 "Whether as the result of ground motion analysis 19 techniques developed subsequent to issuance of the construction 20 permit or data gathered from earthquakes which occurred O
21 subsequent to issuance of the construction permit, the seismic 22 design basis for SONGS 2& 3 is inadequate to protect the 23 public health and safety."
g 24 25 26 O
l O
~
1 TESTIMONY OF DR. STEWART SMITH
( 'T 2
Q.
Would you please state your name?
O 3
A.
Stewart W.
Smith.
4 Q.
Are you the same Stewart W.
Smith who appeared earlier 5
in this proceeding?
O 1
6 A.
Yas.
7 Q.
What is the purpose of your testimony in this portion of 8
the proceedings?
9 A.
One of the questions presented is whether the 10 development of ground motion analysis techniques 11 subsequent to issuance of the construction permit, or 12 data gathered from earthquakes which occurred subsequent 13 to issuance of the construction permits show the seismic 14 design basis of SONGS 2 & 3 to be inadequate.
My 15 testimony discusses seismological advances in the 16 understanding of strong ground motion, both in the 17 g
amounts of available data and in the sophistication of 18 theoretical modelling of earthquake sources.
19 Q.
It has been noted in this proceeding that the numecc of 20 close-in strong ground motion records has approximately nL) 21 doubled in the last eight years.
Have these data 22 affected your conclusions regarding peak ground 23 g
accelerations close-in to large earthquakes?
24 A.
The principal effect has been to provide a solid 25 statistical basis for the view that peak ground 26 acceleration " saturates" or reaches a limiting value O
l O
l 1
almost independent of earthquake magnitude in the
?
('~.)
2 near-field distances (less than 10 km) of a fault 10 3
rupture.
A secondary effect has been to demonstrate 4
that focussing effects on peak acceleration are 5
substantially less than would have been predicted by the lO l
6 kinds of simple rupture models that were available to us I
f 7
at the outset of this project.
Other results are still l
I 8
emerging concerning the effects of site geological lO 9
conditions, fault type, and building size on the 10 acceleration and spectral data now available.
With the 11 increased amount of data, our confidence in prediction 12 of ground motion has increased correspendingly.
There i
13 are still some gaps in strong motion data, for example 14 g
there are no records close to a great earthquake 15 (M
>8).
I believe it will take many years, perhaps i
s 16 into the next century, before we have an extensive data 17 g
base of large earthquakes in a variety of environments.
l 18 The lack of a complete data base however does not 19 prevent making a conservative ground motion prediction 20 g
at any given point in time.
This is precisely what has 21 been done for SONGS 2 & 3, and we have compared our 22 earlier predictions based on the data available in 1972 22 g
with the more than doubled number of strong ground 24 motion records available today.
The conclusion i s that 25 the original design basis remains a conservative one in 26 the light of both improved data and analysis procedures.
2 0
O 1
Q.
What has been the impact of improved theoretical 2
understanding of earthquake rupture phenomena on your O
3 conclusions regarding peak ground acceleration close-in 4
to large earthquakes?
5 A.
This is a field which is still developing rapidly.
Each O
6 new earthquake provides new data to test ideas about the 7
earthquake rupture process.
I believe that the improved 8
capability to calculate wave propagation effects through l
C) 9 realistic earth models, which in turn has permited a 10 better isolation of the actual source effects on ground 11 motion, has been the key to improved understanding.
One O
12 important result that has emerged concerns the basic 13 incoherence of the rupture propagation phenomenon, at l
l 14 least when viewed at high frequency.
The calculational 15 capability that now exists together with the new data 16 has demonstrated that the rupture process cannot be a 17 O
very coherent one otherwise extremely large i
18 accelerations would have been observed during a number 19 of recent earthquakes.
Focussing, a phenomenon of
(
20 considerable concern eight years ago, has proved to be J
21 of limited importance for peak ground accelerations.
22 Further refinements in our understanding of the rupture 23 process and the effects of wave propagation will
,J 24 undoubtedly take place in the future.
With such 25 improved understanding, I would also look for 26
///
2 O
lC 1
reductions in the levels of conservatism necessary in
(/
2 specifying ground motion criteria.
.O l
3 Q.
Has the publication of USGS81-365 had an impact on your i
4 assessment of the conservatism of the SONGS 2 & 3 l
5 seismic design basis?
I 6
A.
No.
An alternative statistical analysis of essentially i
7 the same data had already been completed by TERA 8
Corporation (Exhibit LHW-1, described in the testimony nv 9
of Lawrence Wight) which demonstrated that predicted 10 accelerations for an (Ms 7) earthquake at a distance of 11 g
8.0 km, would be less than those assumed for the 12 original design grcund motion at SONGS 2 & 3.
13 Q.
How can two statistical analyses of essentially the same 14 gggg data yield such different results?
15 A.
In the distance-magnitude ranges where results are 16 strongly controlled by data, there should not be any 17 g
significant difference between the results, and this is 18 in fact the case.
I am speaking of distances beyond 25 19 km and earthquakes less than magnitude 7.
On the other 20 hand, in the range where there are few data, the results 3
21 depend strongly on the model assumptions made in the 22 analysis.
This is the area, less than 10 km in distance 23 for example, where the real differences between these g
24 two statistical analyses appear.
25 Q.
can you describe the differences between USGS81-365 and 26 Exhibit LHW-l?
4 0
O 1
A.
Yes, the principal differences in approach are in the i
)
2 analysis model, the data set, and the weighting O
3 procedure.
Contrasts between the two approaches are 4
summarized briefly below:
5 (1)
Analysis Model:
The basic constreint in the 81-365 6
model is that all earthquakes have the same shape 7
for their distance attenuation curve, irrespective 8
of magnitude.
At several places in 81-365 there 9
are statements that suggest the authors recognize 10 that the bulk of the seismological evidence points 11 toward a different conclusion (pp.
6, 15).
They 12 argue, however, that the data that they have used 13 does not seem to require a magnitude dependent 14 sttenuation curve.
This is partly due to the fact 15 that their selection of data tends to obscure the j
16 magnitude dependence.
More importantly, they 17 O
apparently did not actually test any other models I
18 to see if they might be more appropriate.
Using 19 the tests described in LHW-1 for magnitude 20 dependen e, with the 81-365 data set, shows that O
21 magnitude saturation effects are clearly 22 significant.
Taken together with the 81-365 l
23 g
weighting system, the procedure they have used 24 forces the prediction of peak motion close to large 25 earthquakes to be dominated by more numerous data 26 from larger distances.
Removing the constraint 5
O
s-oh 1
that all magnitude earthquakes have the same shape
('n, kJ 2
for the distance attenuation curve and emphasizing O
3 the more important close-in data, as was done in 4
Exhibit LHW-1, produces an improved statistical fit 5
to the data, and thus is a more useful and reliable O
6 result.
7 (2)
Data Set:
USGS81-365 includes data from shallow 8
soil sites, large distances, and some low O
9 accelerations.
This has the net effect of 10 significantly increasing the scatter in any 11 statistical result.
No new data for large 0
12 earthquakes at close distances is presented.
In 13 fact, there are fewer data in the range 0-10 km, 14 which is of importance to the case at hand, than 15 are presented in LHW-1.
The inclusion of the kind 16 of data first mentioned above seems to me to be an 17 unnecessary degradation of the data base.
For u
18 example, including low acceleration records at 19 large distance, which really have no relevance to 20 the prediction for a large earthquake nearby, 21
^
increases the standard deviation of such 22 predictions without really adding any useful 23 information.
O 24 (3)
Weighting Procedure:
The authors of the USGS 25 report claim that their data blocking procedure has 26 the effect of letting each earthquake contribute 6
O E
O 1
equally to the shape of the magnitude-acceleration
)'
i 2
curve, and each recording to contribute equally to O
3 the shape of the acceleration-distance curve.
In 4
my view, this is inferior to the procedure used by i
5 both IMI-l (described in the testimony of Dr.
I. M.
6 Idriss) and LHW-1, in which data were weighted by 7
earthquake and by distance.
It is important to 8
insure that no single earthquake dominates the v
9 prediction, but it is equally important to make 10 sure that the numerous data available at large 11 distances do not dominate the few but very no 12 important data that are available for distances 13 less than 10 km.
The USGS procedure fails to 14 provide this important feature.
15 Q.
If the analyst's choice of a statistical regression 16 model has such c; important effect on the results 17 obtained, on what basis should that model be chosen?
O 18 A.
There are two types of considerations that should go 19 into model selection.
First, the regression model must 20 O
reflect as much of the physics of the process as is 21 known, and second, it should provide enough flexibility 22 so that the results are primarily controlled by the data 23 insofar as that is possible, and not by the model g
24 assumptions.
25
///
26
///
7 O
O 1
Q.
Do you believe the model used in USGS81-365 is an es
( >)
2 appropriate representation for near field accelerations O
3 from a large earthquake?
4 4
A.
No.
A key element of the USGS81-365 model is the 5
assumption that attenuation can be physically
-O 6
represented by spherical spreading and anelastic 7
absorption.
Such a model is conceptually the same as a 8
point source model and probably works fairly well at O
9 distances in excess of about 40-50 km.
However, this 10 model is inappropriate to use at distances less than 11 about 40 km, where the finite dimension of faulting 12 begins to influence the distribution of peak ground 13 motion within the geometric near-field of the rupturing 14 fault.
At closer distances to large faults, theory and 15 observation have led a number of researchers to propose 16 different physical models of attenuation in the 17 near-field.
Such models recognize the role played by
. %)
18 the presence of an extended rupture surface.
In 19 essence, these models share two properties:
- 1) that 20 attenuation of peak ground motion with distance is a 21 function of magnitude, so that there is a 22 chape-dependence on source size; and 2) that the peak 23 motion values at different magnitudes approach one 24 another in the near-field, so that there is a near-field 25 magnitude saturation effect.
26
///
ll 8
O
O 1
Q.
What is your opinion regarding the applicability of the (m
k-2 results of USGS81-365 to the SONGS 2 & 3 site?
O 3
A.
I agree with the authors of that report, in their 4
statement that "For distances of less than 40 km from 5
earthquakes with M greater than 6.6, the prediction 6
equations are unconstrained by data and the results 7
should be treated with caution".
The net effect of 8
their data selection, model assumed and weighting O,
9 procedure is to produce a result that is a less than 10 optimum representation of near-field ground motion.
I 11 do not believe it is applicable close to large magnitude 12 earthquakes and thus is not appropriate for a M 7
s 13 earthquake at an 8 km distance such as at SONGS.
4lh 15 16 O
18 19 20
,a 21 22 23
,s 24 25 h
9 O
O 1
TESTIMONY OF DR. GERALD A.
FRAZIER O
2 Q.
Would you please state your name?
O 3
A.
Dr. Gerald A.
Frazier 4
Q.
Are you the same Dr. Gerald A.
Frazier who appeared as a 5
witness earlier in this proceeding?
6 A.
Yes.
7 Q.
What is the purpose of your testimony in this portion of 8
this proceeding?
9 A.
Another issue in this proceeding is whether as a result 10 of ground motion analysis techniques developed 11 subsequent to issuance of the construction permit or 12 data gathered from earthquakes which occurred subsequent 13 to issuance of the construction permit, the seismic 14 gg design bases for SONGS 2 & 3 is inadequate to protect 15 the public health and safety.
My testimony in this 16 portion of the proceeding discusses theoretical 17 O
devel pments in ground motion analysis techniques which 18 have occurred since the issuance of the construction 19 permit.
I will also be discussing earthquakes that have 20 O
occurred subsequent to the issuance of the construction 21 permit in the areas of Imperial Valley, California and 22 Victoria, Mexico.
23 O
Q.
How have theoretical considerations of earthquake ground 24 motions changed since issuance of the conetruction 25 permit in 1973?
2e fff O()
O
1 g
w/
1 A.
The principal advancement has been in our ability to
,(
2 implement earlier theories on large digital computers so
)
O 3
as to appraise the effects of various hypothesized 4
conditions.
This work has been directed toward the 5
understanding of two physical processes:
O 6
(1)
The fracture in the earth that produces the 7
observed earthquake disturbance, and 8
(2)
The propagation of seismic waves from the fracture O
9 in the carth to the recording site.
10 The fracture problem is being examined both in a 11
" forward" manner and in an " inverse" manner.
In the O
12 forward approach, methods are developed to model the 13 evolution of a shear crack from its initial fracture 14 through its growth (or spreading phase) to its eventual 15 termination.
Mathematical solutions are restricted to 16 highly idealized representations of the crack problem 17 because of its inherent complexity.
Nevertheless, the 18 limited number of solutions available provide useful 19 information on the speed with which the fracture 20 spreads, the abrupt onset of slip in the vicinity of the 21 lead edge of the crack, and the relationship between 22 fault offset and changes in stress for various sizes of 23 fracture.
These idealized mathematical models have
%J 24 proven useful for explaining the low-frequency (less 25 than about 1 Hz) properties of earthquake fracture.
At 26 higher frequencies, the fracture-induced waves become 2
0
O 1
strongly influenced by fluctuations in the spreading
(
)
2
- crack caused by spatial variations in the subsurface O
3 materials and their state of stress prior to the 4
More precisely, the frequency range of 5
validity for these simplified theories depends on the O
6 ultimate size of the earthquake fracture, but, for the 7
purpose of these proceedings, 1 Hz represents a suitable 8
demarcation.
O 9
Numerical precedures have been developed to 10 simulate the effects of heterogeneities (spatial 11 irregularities) in the earth.
To my knowledge, the 12 recent study performed for the NRC (sponsored by NASA) 13 by Dr. Steven Day ("Three-Dimensional Finite Difference 14 g
simulation of Fault Dynamics", Systems, Science and 15 Software Report SSS-R-80-4295, December 1979) represents 16 the most advanced (forward) treatment of spontaneous 17 earthquake fracture achieved to date.
Day used a 18 numerical method termed " finite difference" to reveal 19 information on how local variations in the pre-existing 20 state of stress causes the rate of crack growth to speed
,sv 21 up and slow down.
His studies also reveal important 22 information about the time evolution of slip at typical 23 points on the fracture surface to assist in evaluating 7,a 24 the abrupt snap that occurs as the crack extends into 25 previously unfractured material.
26
///
n,,
s s
_/
3
'O
O 1
More studies of this type are needed to investigate (7,)
2 the range of likely fracture conditions thought to occur O
3 in the earth.
The un ertainty is large.
For example, 4
the presence or absence of subsurface water can alter 5
failure stresses by a factor of ten, which I believe is 6
roughly the range of uncertainty in the amount of stress 7
residing in the earth prior to an earthquake.
8 Furthermore, even if the failure stress of fault gouge
,u 9
(highly fractured or pulverized rock) deep in the earth 10 were more accurately known, current knowledge fails to 11 reveal whether the average stress resides at 10% or 90%
O 12 of the stress required for failure.
Low values of 13 initial stress (10% of the failure stress) would be 14 ggg appropriate for earthquakes if the cracks were closely 15 contained to the major surface of rupture with 16 relatively little energy lost in the process of 17 O
extending the fracture.
Day has investigated the other 18 extreme, namely high fracture energy and high initial 19 stress (90% of the failure stress).
Consequently, the 20 O
reliability of his result for modelling actual 21 earthquake fracture has not yet been established.
22 Furthermore, comparable uncertainty exists in the 23 g
estimation of how much load the crack surface is able to l
24 carry while sliding is in progress.
Day considered only 25 about a 10% reduction in stress at the onset of 26 fracture; whereas, one could just as easily argue for
)
%.j' 4
0
- O 1
nearly total release of stress during fracture.
The G
'D 2
energy released by earthquakes is sufficient to melt a 10 3
layer of rock along the fault face, which would act as a 4
lubricant for freeing stresses during sliding.
l 5
Because of the range of uncertainties that exist in lO 6
the fracture physics of earthquakes, especially 7
regarding the causes of fluctuations in the fracture, 8
simulations of earthquake fracture (forward methods) 9 must be considered to be highly idealized.
I believe 10 that such theory can best be used as a guide in the 11 gross characterization of earthquaka fractures; actual
(,/
12 earthquakes have not yet revealed all of their secrets.
13 Q.
What alternative is there to the forward approach of 14 simulating the earthquake fracture?
l 15 A.
The inverse approach noted above.
Using the inverse 16 approach, researchers rely on observational data 17 g
(principally ground motion recordings) to determine what 18 happened in the fracture zone.
Unfortunately, as with 19 the forward problem, the level of difficulty in 20 O
d termining fracture behavior increases with increasing 21 frequency until the detailed evolution of an earthquake 22 fracture becomes obscured.
Presently, resolution is 23 G3 limited to frequencies below about 1 Hz which 24 corresponds to spatial fluctuations about 3 km or larger 25 in size,
///
O 5
O
O 1
Dr. Randy Apsel and Mr. Allen Olson (" Numerical
\\ /
'~'
2 Inversion for Localized Zones of Energy Release",
- O 3
Abstract in the Seismological Society of America Annual 4
Meeting, March 23-25, 1981, Berkeley, California) have 5
recently presented, what is to my knowledge, the most
_U 6
detailed interpretation of an earthquake fracture yet 7
achieved.
They relied on more than a dozen near-field 8
recordings of the 1979 Imperial Valley Earthquake to w/
9 invert for the progression of fault slip that best 10 explains the low frequency portion (below 1 Hz) of these 11 recordings.
Incidently, they found fracture g
12 irregularities even in the low frequency components of 13 fracture.
14 Q.
If both the forward methods and the inverse methods have ggg 15 problems at high frequencies, what is the basis for 16 modeling the high frequencies of earthquakes?
17 A.
Details of the fracture sequence have not been nLJ 18 determined from high frequency recordings.
Inversion of 19 high frequency data does, however, provide gross 20 O
properties of the fracture process such as the average 21 production of high-frequency seismic waves over the 22 frequency range of interest.
Accordingly, the most 23 g
reliable characterization of earthquake fracture is 24 obtained by adjusting results obtained from simulations 25 of idealized cracks to produce results consistent with 26 g
recorded motions.
When the two approaches, forward and 6
O
O 1
inverse, are considered jointly, the inverse approach I
i 2
must take precedence to ensure realistic simulation of O
3 recorded earthquake motions.
4 Q.
What approach did you use in the modeling study set 5
forth in Exhibits GAF-1, 2,
3 & 4?
O 6
A.
I used the combined approach of forward and inverse 7
methods.
Forward methods were used to set the 8
mathematical representation of fractures; while inverse 9
methods, which relied on earthquake recordings, were 10 used to set the parameters in this representation.
We 11 cannot resolve the detailed properties of earthquake u
12 fracture, so we have provided the least complicated 13 characterization of the fracture process that is 14 consistent with recorded data.
15 Let me elaborate on this often overlooked step.
16 The simplest characterization of earthquake fracture 17 involves a coherent fracture that spreads without 18 fluctuations.
This characterization causes about ten 19 times more high frequency waves in the direction of the 20 expanding rupture than has been observed in g
21 acceleratians recorded for of past earthquakes.
This 22 characterization does, however, provide reasonable 23 results for frequencies below about 1 Hz.
O 24 To subdue the excessive focusing of high-frequency 25 waves in accordance with observations, irregularities 26 have been introduced into the simulated fracture
,3
,ymv) 0 7
O
i O
1 process.
We have accomplished this through the lf3)
~
2 introduction of random irregularities in such source O
3 properties as the rupture speed, the crack orientation, 4
and the slip direction.
Each earthquake simulation is 5
repeated several times to appraise the range of effects m,
6 provided by the randomly prescribed irregularities.
7 This is the most realistic approach yet achieved for 8
modeling broad-band earthquake motions within 20 km of mu 9
the-causative fracture.
10 Q.
You earlier stated that earthquake theory has been 11 O
directed toward the understanding of two physical 12 processes:
earthquake fracture and wave propagation.
13 How do theoretical advercements in the taderstanding of
.ggg wave propagation relate to ground motions close to 14 15 earthquakes?
16 A.
Once the fracture process is suitably characterized, the 17 seismi waves thus produced must travel through a O
18 realistic representation of the subsurface geology in 19 order to simulate ground shaking at the earth's 20 surface.
.Secause the transmission velocities of rocks O
21 tend to increase with increasing depth, the trajectory 22 of seismic waves bends (refracts) toward the earth's 23 surface.
This effect is larger in some geologic regions g
24 than in others.
For example, the shallowest 6 25 kilometers of the Imperial Valley region of Southern 26 California bend seismic waves toward the earth's surface g
8 O
- O 1
more dramatically than the geologic region surrounding b) 2 the SONGS site.
10 3
Also, some of the seismic waves are trapped in the
(
4 surface layers and forced to travel outward from the S
frac ture zone adj acent to the surface.
The consequence O
6 is that we experience disproportionately strong shaking 7
as compared to motions deep within the earth.
Wave 8
theory has only recently been developed for simulating 9
the complete myriad of (high and low frequency) waves 10 that exist in an idealized, layered representation of 11 the earth.
To my knowledge, the computer method mu 12 (PROSE), used in the modeling study sponsored by me, is 13 the best method currently available for modeling ground 14 motions near to large earthquakes.
15 Q.
Are there other recent theoretical developments that 16 relate to the predictions of earthquake motions at the 17 SONGS site?
O 18 A.
Yes, considerable progress has been made in the 19 theoretical understanding of earthquakes in recent 20 g
years.
The need to ensure conservative estimates of 21 potential ground shaking for man-made structures is 22 stimulating considerable research.
The " burden of 23 proof" requisite for nuclear power plants near " capable"
,,u) 24 faults is significantly influencing the quantification 25 of earthquake effects.
Advancements to our 26 understanding of earthquakes are occuring rapidly with 9
O
- O 1
contributions from many researchers.
I have described O
2 only a small portion of this recent work by categorizing
- O 3
the most relevant research areas and describing the 4
leading edge of knowledge in these areas.
5 I want to emphasize that these descriptions of 6
curren; developments are based on my knowledge and 7
experiences; other researchers might prefer a somewhat 8
different emphasis.
g 9
Q.
What has been learned from recordings of earthquakes 10 since the construction permit was issued in 1973?
11 A.
O The number of recordings close to potentially damaging 12 earthquakes has approximately doubled during thic span 13 of eight years.
The principal consequence of these new 14
(}
data has been to confirm earlier predictions.
For a 15 specified distance from an earthquake of known 16 magnitude, peak accelerations are predicted to within 17 33 about 50%.
That is, the median prediction plus 50%
18 encompasses approximately 84% of the recorded peaks.
19 This range of observed acceleration peaks characterizes 20 O
the variations among recordings of a single earthquake 21 about as well as between different eartnquakes.
This 22 suggests that uncertainties in the predicted peak 23
.g acceleration stems principally from two sources:
(1) 24 irregularities in the fracture process and the geologic 25 materials traversed by the earthquake waves, and, (2) 26 g,,
systematic variations in ground motions due to b
10 0
,r.)
1 7,
orientation with respect to the expancing fracture.
If V
2 O,
uncertainties due to unforseen effects of a single 3
earthquake were of first order importance, then I would 4
expect to see greater differences in peak accelerations 5
-)
between different earthquakes than those observed for a 6
single earthquake.
To the contrary, peak accelerations 7
for one earthquake are about as consistent with those 8
g recorded for other earthquakes, of comparable ma9nitude 9
and distance, as they are self consistent.
This 10 reassuring feature has become apparent with the 11 3
significant increase in recordings close to large 12 earthquakes.
13 Q.
Would you describe what was learned from the 1979 ggg 14 Imperial Valley Earthquake?
15 A.
The Imperial Valley Earthquake of October 15, 1979 16 (IV 79) was located on the Imperial Fault in 3
17 approximately the same location as the 1940 Imperial 18 Valley Earthquake, pictured in Figure GAF-E, " Strong 19 motion stations for the 1979 Imperial Valley 3
20 earthquake...."
The IV-79 event generated more strong 21 motion recordings than any other strike-slip earthquake 22 to date.
There were more than 25 three-component 23 l3 recordings in California and about eight additional 24 recordings in Mexico that are of :mecial interest.
25 Recorded values for the largest acceleration at each l
26 station within 22 km of the fault trace are tabulated in l
11 L
O 1
Figure GAF-F, " Calculated peak acceleration values for
(
1 2
the October 15, 1979 Imperial Valley Earthquake compared 3
to the recorded values..."; the acceleration peaks for 4
the vertical and horizontal components are presented n
5 graphically in Figures GAF-G, " Peak' Vertical O,
6 Acceleration Values for the 1979 Imperial Valley 7
Earthquake" and GAF-H, " Peak Horizontal Acceleration 8
Values for the 1979 Imperial Valley Earthquake."
The g
9 surface-wave magnitude for this earthquake has been 10 assigned a value of 6.9 with seismic moments ranging 11 from.8 x 102' to 2.0 x 1025 dyne-cm.
g 12 The instrumentation for IV-79 is sufficiently 13 extensive so as to be able to derive important 14 gggg information regarding the properties of the earthquake 15 fracture, the characteristics of strong ground shaking 16 close to the large fracture, and the attenuation of the 17 g
stre g shaking with increasing distance from the 18 fracture.
Recording stations range from essentially on 19 top of the exposed fracture to distances extending 20 around the world with abundant recordings at all g
21 azimuthal directions from the rupture.
The abundance of 22 near-field recordings provides a challenging test to the 23 predictive capabilities of earthquake models including g
24 those used to provide estimates at SONGS.
25 Some earthquake hypotheses have been inva'idated 26 while others have been confirmed.
The most impu
- nt (y_()
12 O
O 1
f3 concepts that have found added support, if not total
'w.)
2 confirmation, include:
nv 3
(1)
Peak accelerations close to large earthquakes are 4
essentially independent of the earthquake 5
O magnitude.
Earlier recordings close to smaller 6
earthquakes have peak accelerations remarkably 7
similar to those for IV-79.
8 (2)
The effects of focusing on acceleration peaks are g
9 much less important than those obtained from 10 simple, coherent fracture theory.
Prior to IV-79 11 there was a reasonably plausible explanation for O
12 our failure to record unusually high accelerations 13 directly in the earthquake's path due to ggg 14 insufficient spatial coverage.
As seen in GAF-E 15 there is essentially no portion of the fracture 16 left uncovered by recording instruments for IV-79.
O 17 The average horizontal acceleration close to the 18 surface exposure of fracture is less than 0.5g 19 regardless of whether averaged over the nearest one O
20 kilcos 3r, two kilometers, or any other distance.
21 (3)
Peak horizontal accelerations saturate with 22 decreasing distance to the earthquake fracture.
O 23 This effect, illustrated by GAF-H has never been so 24 clearly demonstrated as for IV-79.
25 (4)
The scatter in recorded acceleration peaks from one O-s 26 recording site to another is comparable to the
\\
13 O
n
-Ql 1
scatter in acceleration peaks recorded for several 2
different earthquakes when properly scaled for g
3 magnitude and distance effects.
As explained 4
above, this indicates that dissimilarities in O
5 earthquake fractures are not the major source of 6
uncertainty in predicting ground motion.
The major 7
uncertainty results from whatever mechanism causes O
8 dissimilarities in the recordings for a single 9
earthquake (principally station-correction 10 factors).
The scatter in peak accelerations O
11 recorded for IV-79 (cf. GAF-H), is typical of data 12 scatter for other earthquakes.
13 IV-79 is similar in several important respects to qll 14 the largest earthquake being contemplated along the OZD 15 directly offshore to the SONGS site.
The source 16 mechanism and the magnitude are essentially the same, O
17 namely strike slip and Ms = 7.
Furthermore, the 18 earthquake occurred in the tectonic setting of Southern 19 California with an abundan 3 of records at distances of O
20 interest for SONGS.
21 IV-79 is dissimilar from the postulated largest OZD 22 event in one important respect:
geology.
The O
23 crystalline basement rock at Imperial valley forms a 24 deep valley (about 6 km deep) that has been filled, over 25 recent geologic time, with low-velocity sediments, while Cl-26 the SONGS area can be characterized by a shallov l \\i 1
14 10 1
1
O 1
basement overlain by consolidated higher velocity c
(s_)
2 sedimentary rocks.
The subsurface material velocities r.J 3
at Imperial Valley are contrasted with those at SONGS in 4
Figure GAF-I, " Comparison of the P-wave velocities for 5
SONGS earth structure (from Table 6-1 of Exhibit GAF-2) g 6
and Imperial Valley earth structure (from Table 3-2 of 7
Exhibit GAF-4)."
8 According to both observational data and earthquake g
9 theory, the low-velocity sediments at Imperial Valley 10 are expected to cause an upward bias in recorded 11 velocity peaks and vertical acceleration peaks as g
12 compared to conditions at SONGS.
The upward bias in 13 recorded velocity peaks due to soft sediments has been 14 well observed and documented prior to IV-79 [See, for ggg 15 example, page 18 of USGS-795).
The mechanism by which 16 the sediments at Imperial Valley amplify vertical o
17 accelerations has become understood more recently, and 18 consequently this phenomenon is not so well documented.
19 Q.
Would you elaborate on the amplification of vertical O
20 accelerations recorded for IV-79?
21 A.
Several of the instruments within about 10 km of the 22 exposed fracture recorded larger vertical accelerations o
23 than many earthquake scientists would have expected.
24 Upon examination of these close-in recordings, we find 25 that the unexpectedly high vertical accelerations result g_
26 from P waves (acoustic-like waves that are first to V
15 O
O
<~s 1
arrive) rather than the much more energetic S waves
(
)
\\m/
2 (waves that travel with shearing motions).
The P waves O
3 for IV-79 contain much less total energy than the S 4
waves.
At high frequencies, however, the amplitude of 5
the P waves exceeds that for S waves, and consequently, C
6 large vertical accelerations were recorded.
7 Q.
Is a similar condition likely for SONGS 7 0
8 A.
No.
The slow, gradual increase in material velocity 9
with increasing depth at Imperial Valley, illustrated in 10 GAF-I, causes earthquake waves to bend (refract) more O
11 sharply toward the earth's surface than the more typical 12 velocity profile beneath SONGS, also illustrated in 13 GAF-I.
Consequently, the P waves that radiate
'dll 14 horizontally outward from the fracture surface emerge at 15 the earth's surface closer to their source in the 16 Imperial Valley than at SONGS.
Figure GAF-J, " Emergent O
17 P-waves for horizontal source radiation", illustrates 18 that, for the particular case of P waves originating at 19 a depth of four kilometers, Imperial Valley waves emerge O
20 in about half of the distance required for the SONGS 21 geology.
This implies more energetic (higher amplitude) 22 P waves close to the earthquake fracture in the Imperial O
23 Valley than would be expected for SONGS.
24 The earthquake fracture for IV-79 extends from its 25 surface exposure, through the 6-km-deep sedimentary QS 26 deposit, and into the underlying basement rock to a
%.,)
16 lC
O
(~)
1 depth of about ten kilometers.
The unusually close v
O 2
emergence of P waves in the Imperial Valley occurs for 3
fractures at all depths but particularly for fractures 4
in the deep sedimentary deposit.
At all depths, approximately 50 percent of the earthquake waves radiate O
5 outward from the fracture along an upward trajectory and 6
Because P waves 7
50 percent along a downward trajectory.
O 8
leaving the fracture at a depth of four kilometers along 9
a horizontal rajectory emerge within about six 10 kilometers of the epicenter (GAF-J) 50 percent of all P O
11 waves produced by such a fracture emerge within thin 12 distance.
In contrast, twice as much distance is 13 required at SONGS before 50 percent of the Mlh 14 fracture-induced waves emerge.
15 The percentage of P waves that emerge within other 16 distances from the fracture is presented in Figure 0
17 GAF-K, "P-wave energy emerging within specified 18 epicentral distances for Imperial Valley earth structure compared with SONGS and a homogeneous half-space."
Also 19 n'
20 presented in this figure for the purpose of comparison are results for the idealized case of uniform material 21 22 velocity (homogeneous earth) in which waves follow 23 straight trajectories (no biasing of waves toward the n
24 earth's surface).
Earth properties for SONGS cause P 25 waves to emerge at a distance about midway between that
'/ ~)
26 for IV-79 and that for unbiased, homogeneous earth.
The n
n.e 17 O
O r's 1
results presented in GAF-K are reliable in that small
(/
2 changes in earth properties (within our range of g
3 uncertainty for their true properties) result in almost 4
no modification to these curves.
5 I would not, therefore, expect vertical O
6 accelerations for the postulated earthquake offshore of 7
SONGS to be as unusally large as those recorded for 8
IV-79.
Based on earthquake theory, of the type g
9 described above, I estimate that the vertical 10 acceleration peaks recorded for IV-79 should be reduced O
11 by about 30 percent before comparisons are made with 12 postulated earthquake motions at SONGS.
13 Q.
Have you considered wave properties of the Imperial 14 Valley Earthquake other than the refraction properties 15 described above?
16 A.
Yes.
Additional amplification occurs in the Imperial
- O 17 Valley for vertically emerging waves.
The stiffness and 18 density properties decrease along the path of vertically 19 emerging waves more dramatically in the Imperial Valley O
20 than for SONGS.
To conserve energy, the amplitude of 21 the wave increases as the emerging waves approach the 22 earth's surface, again more dramatically in the Imperial O
23 Valley than for SONGS.
A similar type of amplification 24 occurs in the ocean where the amplitude of waves gets 25 larger as the waves approach the shore.
This upward e
26 bias at Imperial Valley is mitigated to some degree by 18 O
.O 1
material hysteresis.
The unconsolidated sediments in g3 N./
2 the Imperial Valley probably dissipate seismic energy g
3 more readily than the consolidated materials underlying 4
the SONGS site.
This mitigating factor is much more 5
g prevalent for S waves than for P waves.
In -he 6
near-surface sediments at Imperial Valley, S waves above 7
lOHz can be attenuated by a factor of ten within one g
8 kilometer, while P waves of the same frequency are 9
attenuated by only about 20 percent over the same 10 one-kilometer distance.
o 11 Even accounting for the effects of the unique earth 12 structure at Imperial Valley, the large vertical 13 acceleration recorded at Station 6 (1.74 g later dll 14 corrected to 1.52 g) during the 1979 Imperial Valley 15 earthquake poses an enigma.
Station 7, which is about 16 1 km SW of Station 6, recorded a peak vertical O
17 acceleration of.65 g (corrected to.51g), and Station 18 5, which is approximately 3 km NE of Station 6, recorded 19 a peak vertical of.71 g (corrected to.44g).
It is O
20 clear from field observations that the Imperial Fault 21 passes between Station 6 and Station 7 and the Brawley 22 Fault passes between Station 6 and Station 5 It is O
23 important to note that the motion of the wedge-shaped 24 block, upon which Station 6 is located, is downward 25 relative to the adjacent blocks.
.C x 26
///
V I
19 1O
O 1
Boore and Fletcher ("A preliminary study of 7_
! /
2 selected aft'ershocks of the 1979 Imperial Valley u
3 earthquake from digital accelerations and velocity 4
recording",
U.S.G.S.
reprint, 1980) note that recordings 5
of an aftershock which occurred to the south of Station 6
6 and Station 7 indicate that the spectral amplitudes of 7
P waves were considerably higher at Station 6 than at 8
Station 7, for a wide range af frequencies.
Porchells Oc 9
(" Seismic Engineering Program Report, September-December 10 1977",
U.S.G.S.
Circular 762-c, 1978) observes that a 11 q
swarm of earthqakes in 1977 north of the stations show v
12 the peak accelerations recorded near Station 6 to ',e 13 consistently larger than those recorded near Station 7.
14 ggg Thus, the motion at Station 6 is amplified relative to 15 Station 7, basically independent of the direction of 16 incoming events.
17 O
Of significance is the observation that the S waves 18 at Station 6 were delayed by.5 seconds relative to 19 Station 7 (Boore and Fletcher, 1980).
Station 6 is 20 g
underlain by material of lower velocity than is Station 21 7.
The lower velocities may be due to the down-dropping 22 of the wedge-shaped block between the Imperial and 23 g
Drawley Faults upon which Station 6 is located.
This 24 would result in lower velocity sediments beneath Station 25 6 that extend deeper than those beneath either Station 7 26 or 5.
A laterally heterogeneous earth structure, which nN 3
%s 20 0
O 1
exhibits a column of low velocity material beneath the
(,)
['~'
2 vicinity of Station 6, would trap obliquely emerging U
3 waves by refracting the wave toward the region of lower 4
velocities (i.e.,
into the wedge).
This is believed to 5
be the cause of the significant amplification of 6
vertical accelerations recorded at Station 6.
Material 7
attenuation, as noted above, prevented severe 8
amplification of the high frequency S waves, which mu 9
represent the major contributor to the horizontal ground 10 acceleration.
11 Extensive geologic investigations at the San Onofre nv 12 site indicate that similar local amplifications of 13 ground motion should not occur at that site.
Therefore, 14 g gg such unusual recordings as that for Station 6 do not 15 provide suitable analogs for use in evaluating the 16 seismic criteria at SONGS.
17 g
Considering the composite of these differences in 18 geology between Imperial Valley and SONGS, I conclude 19 that horizontal motions for IV-79 are representative of 20 what to expect for a comparably large earthquake on the g
1 l
21 OZD while vertical accelerations would be about 30%
l l
22 lower at 8 km from the OZD than those for IV-79 at a l
23 comparable distance.
g 24 Q.
You previously stated that you had modeled IV-79.
Would 25 you describe the results of this study?
26
///
21 lC
O 1
A.
Yes.
I will describe the major results of our computer 73V 2
simulation of IV-79 which are presented in greater
,a 3
detail in GAF-4.
Subsequent to the completion of 4
Supplement III in August of 1980, additional recording 5
stations were modeled.
These supplementary results are g
6 included along with the results from Supplement III in 7
Figure GAF-F, G,
and H.
Additionally, detailed 8
g inversion studies have been performed to determine the 9
low-frequency portion of the fracture for IV-79, as 10 previously described (reference:
Apsel and Olsen, g
" Numerical Inversion.?or Localized Zones of Energy 11 12 Release," supra).
My discussion herein pertains to 13 higher frequencies, i.e.,
frequencies greater than 1 Hz.
14 The fracture model developed in Supplement I ggg 15 (GAF-2) was tested against near-field recordings for 16 IV-79.
These simulations demonstrate the suitability of g
17 Supplement I model for predicting horizontal 18 accelerations in the distance range of interest for 19 SONGS (Chapter 4 of GAF-4); however, this model was O
20 found to underestimate irregularities in the fracture 21 process for IV-79.
22 Consequently, additional randomness was o
23 implemented.
This modified characterization of 24 earthquake fracture is denoted "IV-79 refined model".
25 Further simulations of IV-79 were produced using the 26 IV-79 refined model as described in Chapter 5 of GAF-4.
22 O
D 1
It is these results, plus subsequent results that were
)
2 performed for yet additional recordings stations, that D
3 are presented in GAF-F, G,
and H.
The computed vertical 4
accelerations lie well within the range of recorded 5
vertical accelerations for both peak values and spectral 6
values.
The computed horizontal accelerations generally 7
exceed recorded horizontal acclerations within 3 km of 8
g the surface fracture, indicating exaggerated focussing 9
in the computer simulation.
Good agreement is achieved 10 for distances of interest for SONGS.
The computed 11 g
horizontal accelerations are deficient in high-frequency 12 energy for several stations at distances greater than 10 13 km from the surface fracture.
Apparently, the
"')
14 dissipation properties for S waves are not adequately v
15 modeled.
Unfortunately,
-o better information exists at 16 this time for assigning these dissipation parameters.
17 p
The compressional-wave material attenuation in our model 18 gives the appropriate amount of diesipation for P-wave 19 energy as evidenced by the match with the recorded 20 vertical accelerations as a function of distance.
g 21 Q.
Based on your studies of current earthquake data and 22 theory, have you drawn any conclusions about the seismic 23 criteria used for SONGS?
y 24 A.
Yes.
Both earthquake theory and earthquake recordings 25 support my conclusion that the DBE used for SONGS is 26 conservative with respect to the largest postulated 23 D
W O
1 earthquake on the OZD.
The consistency of earthquake 2
re rdings minimizes the possibility of anomolous future O
3 earthquakes.
4 5
O 6
7 8
0 9
10 O
11 12 13 14 15 16
- O 17 18 19 O
20 21 22 0
23 24 25 26 l
24 O
O l
ol l
Calipatria pV O
_ 33 00'
. Brawley Superstition Airport Mountain e #1 e
0 P rachute I ha
- #2
/
- 3 Imperial Fault
- 6 e#5*#4 e
, f8
- Holtville e Plaster City DA
,10 EPICENTER 1940
_ 32 45' e
Me land
- 11 O
, p12 Bonds Corner e,__~,_-
- 13 Calexico
~
e U.S A-
_ ~
' ~~~ _
, Aeropuerto EPICENTER 1979 gexico grarius e
M Mexicali g
Compuertas S A Hop e
e Cucapah O
- 32 30' Chihuahua e
Cerro Prieto g
g g
j 0
10 20 30 e
(km)
Delta o
O 115*45' 115 30' 115 15'
$ictoria i
l I
~
Figure GAF-E.
Strong motion stations for the 1979 Imperial Valley earthquake.
Imperial fault trace is for the 1940 Imperial Valley earthquake.
!O
(
.O Distance
- Component Observed Calculated Station (km)
(deg)**
(9)
(9)
U O
6
.7 Up 1.52
.686 230
.433 1.12 140
.348 1.05 7
.8 Up
.514
.545 230
.463
.'743 O
140
.333
.7 31 5
1.1 Up
.441
.618 230
.375
.767 140
.530
.819 O
Aeropuerto 1.3 Up
.179
.436 45
. 31 6
.674 315
.240
.782 Agrarius 2.3 Up
.784
.251 03
.339
.390 0d 273
.196
.295 Bond's 3.0 Up
.355
.399 Corner 230
.786
.584 140
.587
.477 8
4.3 Up
.354 230
.467
.451 140
.610
.437 4
4.8 Up
.203
.370 n
230
.357,
.340
~
140
.494
.447 Differential 5.7 Up
.659
.324 Array 360
.487~
.401 270
.352
.288 O
Hol tville 7.4 Up
.228
.266 315
.217
.214 225
.251
.218
~* Closest distance to model fault trace.
- Measured clockwise from North for horizontal components.
Figure GAF-F.
Calculated peak acceleration values for the October 15, 1979 Imperial Valley Earthquake compared to the recorded values.
Values for stations in U.S.
were obtained from U.S.G.S.i while values for stations in Mexico were ob-tained from IGPP; calculated values were obtained using the Supplement III Model from Exhibit GAF-4.
O
O Figure GAF-F (continued)
(^)
Distance
- Component Observed Calculated Station (km)
(deg)**
(g)
(g)
O liexic ali 7.6 Up
.332
.166 03
.311
.271 273
.459
.228 O
10 9.3 Up
.105
.208 50
.172
.206 320
.226
.201 Brawley 9.4 Up
.153
.114 315
.221
.124 0
225
.165
.167 3
9.5 Up
.132
.172 230
.223
.204 140
.267
.187 O
Calexico 9.9 Up
.183
.1 51 315
.201
.169 225
.275
.236 2
10.5 Up
.110
.128 230
.414
.156 140
.316
.209 Cucapah 11.9 Up
.124
.056 03
.0 81 273
.300
.078 O
11 12.7 Up
.140
.151 230
.382
.176 140
.363'
.155 1
16.1 Up
.071
.103 230
.143
.109 O
140
.145
.089 12 17.3 Up
.067
.100 230
.116
.111 140
.142
.082 O
13 22.2 Up
.043
.074 230
.137
.083 140
.11.7
.071 Closest distance to model fault trace.
- Measured clockwise from North for horizontal components.
O
O A
O PEAK VERTICAL ACCELERATION VALUES FOR THE 1979'If1PERIAL VALLEY EARTHOUAKE g
g
\\
A
~
o
~s A
N N
N g
N N
.O i
N
\\
5 N
~
~
p N
~
s O
g A
s g
s A
\\
N d
\\
A i
u g
s N
\\
N N
h
\\
A,^
p N
\\
5-N**
\\
O x
g
\\
g
-O s
g-
\\ '*
\\-
Calculated (Supp. III Model) g g_
A IV-79 Data
\\
\\
iO Mean Regressed Curve
^
\\
~
Mean i la using IV-79 data A = 30.7
\\
\\
i i i i i i
i i
i i i ii i i
\\,
lg l.
10.
DISTANCE (KM)
I l
F i gu re G A F-G.
Calculated peak vertical acceleration values for the October 15, 1979 Imperial Valley Earthquake compared to the recorded values as well as to the regressed mean and mean plus one-standard-deviation curves of the recorded values.
The distances used correspond to the O
closest horizontal distance of the station to the model fault trace.
I I I I I I
I i
l l
1 1 I I I
I v
PEAK HORIZONTAL ACCELERATION VALUES
(}.
- FOR THE 1979 IMPERIAL VALLEY EARTHQUAKE O
l lO
^
l A
O
-. s A
s g
A N(A j.
A s
A e
N A
N O
S A
A
\\
A A A
=
A A
\\
.,b 2
A A\\
'4 ag s q 's g
4 N
A N
\\
s g
4' g
O wd i
4A' N
N.
s yN
\\
g t
N z
i 4
R-
\\^
N E
hA
\\
e~
.* \\
s j
\\
- g O
t g
g
. Calculated (Supp. III Model)
A_
A IV-79 Data
\\
O A
Mean Regressed Curve (R+20)1*75 Mean + 1 using IV-79 ; data gives A = 95.0
~
l I I I I I
I I
l l
l l 1 1 I
O 1.
10.
DISTANCE (KM)
Figure GAF-H.
Calculated peak horizontal acceleration values for the October 15, 1979 Imperial Valley Earthquake compared to the recorded values as well as to the regressed mean and mean plus one-standard-deviation curves of the recorded values.
The distances used correspond to the cl sest horizontal distance of the station to O
the model fault trace.
o P-WAVE VELOCITY (KM/SEC)
O 0
1 2
3 4
5 6
7 8
9 g
1 Q
l I
l l
l I
1 SONGS 3
IMPERIAL
~
VALLEY
~
'o 4
~l l
5 o
6
^zx 7
c W
4 S
8 m
~
g o
10 -
11 -
o 12 -
SONGS IV
~
i 13 -
o i
i i
i i
i i
i yy Figure GAF-I.
Comparison of the P-wave velocities for l
SONGS earth structure (from Table 6-1 of l
Exhibit GAF-2) and Imperial Valley earth structure (from Table 3-2 of Exhibit G A F-4).
l o
O O
.O O
EPICENTRAL DISTANCE (KM) 0 2
4 6
8 10 12 >
O i
i l
l IMPERIAL 2
VALLEY O
^r3 SAN z
ONOFRE s
~
N P
l
.O V
' O l
Figure GAF-J.
Emergent P-waves for horizontal source radiation.
O l-l
'O
i
'Q O
S0uaCE DeerH = 2 xM 100 0
SONGS g 80
[ 60 IMPERIAL O
o VALLEY H*
40 HOMOGENEOUS EARTH 58 20 g
i i
i i
1 2
5 10 20 50 100 EPICENTRAL DISTANCE (KM) 4 SOURCE DEPTH = 4 KM 100 m
O NS SONGS g
R 60 IMPERIAL ao VALLEY g
O 5 40 M
W HOMOGENEOUS EARTH S 20 O
I 1
2 5
10 20 50 100 EPICENTRAL DIST.ANCE (KM)
Figu e GAF-K.
P-wave energy emerging within specified epicentral distances for Imperial Valley earth structure compared with SONGS and a homogeneous half-space.
O
O 1
TESTIMONY OF DR. SHAWN BIEHLER
/_T 2
Q.
Would you please state your name?
3 A.
Dr. Shawn Biehler.
4 Q.
By whom are you presently employed?
5 A.
I am an A:sociate Professor of Geophysics at the g
6 University of California, Riverside, and an Associate 7
Research Scientist at The Institute of Geophysics and 8
g Planetary Physics of the University of California.
In 9
addition I am a private consulting geop).ycist.
10 Q.
In what manner are you associated with the Applicants in 11 this proceeding?
g 12 A.
Over the past six years I hiv conducted various applied 13 geophysical investigations for applicants including a 14 historical and present seismicity study and qgg 15 investigations on the regional and local tectonic 16 structure of the San Juan Capistrano arer.
3 17 Q.
Would you please describe your formal trainin7 in 18 applied geophysics?
19 A.
I graduated cum laude from Princeton University :.n 1958 g
20 with a Bachelor of Science in Geological Engineerit:g and 21 received a Masters and PhD in geophysics from the 22 California Institute of Technology in 1961 and 1964, O
23 respectively.
24 Q.
What professional positions have you had in the area of 25 applied geophysics?
O,,
26
///
( )
Q,/
i I
O 1
A.
I have been an active geophysical consultant to several
\\
2 private corporations and government agencies concerned O
3 with the siting of nuclear power plants, regional and 4
local tectonic structure, ground water studies, and 5
geothermal development.
With regard to nuclear plants, 6
I have been involved in geotechnical investigation for 7
proposed sites in the Mojave Desert and San Joaquin 8
Valley, California, Yuma area of Arizona, and the State 9
of Washington.
I have also been involved in applied 10 geophysical studies of various water agencies and the 11 Guam Environmental Protection Agency.
,J 12 Q.
Have you been associated with any educational 13 institutions?
14 A.
From 1964-15 66 I was a post-doctoral research assistant qg 15 at the Seismological Laboratory of the Cr.lifornia 16 Institute of Technology, Pasadena, California.
From 17 1966-1970 I was an Assistant Professor of Geophysics at 3
18 the Massachusetts Institute of Technology, Cambridge, 19 Mass.
From 1970 to present I have been an Associate 20 3
Professor of Geophysics in the Earth Science Department 21 and an Associate Research Scientist in the Institute of 22 Geophysics and Planetary Physics at the University of 23 California in Riverside.
3 24 Q.
Do you h 'd any professional registrations in the State 25 of Califor.aa or any other state?
26
///
gu t
m s' 2
3
O 1
A.
Yes, I am a registered geophysicist No. 411 in the State 7y Q) 2 of California.
g 3
Q.
What are your pertinent professional or organizational 4
memberships?
5 A.
Geological Society of America; American Geophysical g
6 Union; Society of Exploration Geophysistc; European 7
Association of Exploration Geophysicists; Seismological 8
g Society of America; Society of Sigma Xi; American 9
Association of Advancement of Science.
10 Q.
Have you written or published articles in the field of Il "PPli*d 98 P Y8iC87 h
O 12 A.
Yes, I have authored or co-authored numerous papers and 13 reports dealing with applied geophysics.
A list of ggg) 14 these published and technical reports is appended hereto.
15 Q.
Have you served with formally-organized groups concerned 16 wholly or in part with mattters of seismic safety?
O 17 A.
Yes, I have been an advisor to the City and County of 18 Riverside in the review of the seismic safety element.
19 I have contributed to the State of California seismic
'O.
20 hazards proposal.
I have been a member of the panel of 21 seismic hazards associated with geothermal development 1
22 in the Imperial Valley.
LO 23 Q.
on which projects have you been retained as an axpert 24 consultant in geophysics?
25 A.
I have been retained by a number of governmental l
26 agencies and private corporations over the past 15 years 1
l 3
l 10
O 1
as an expert consultant in the area of applied V
2 geophysics.
A list of these projects is appended hereto.
no 3
Q.
Have you presented expert opinions or tastimony?
4 A.
Yes, I presented expert opinion to the ACRS subcommittee g
in this proceeding.
I have also made presentations as 5
6 an expert before the NRC pertinent to the then proposed 7
Sundesert nuclear power plant site.
8 Q.
What is the purpose of your testimony in this proceeding?
O 9
A.
One of the issues in this proceeding is whether as the 10 result of ground motion analysis techniques developed O
11 subsequent to issuance of the construction permit or 12 data gathered from earthquakes which occurred subsequent 13 to issuance of the construction permit, the seismic ggg 14 design basis for SONGS 2 and 3 is inadequate to protect 15 the public health and safety.
In this regard, the 16 purpose of my testimony As to assess the significance of O
17 two small earthquakes which occurred in Trabuco Canyon 18 in 1975 and a cluster of microseismic events that also 19 occurred in Trabuco Canyon in 1977.
l
.O 20 Q.
Could you generally describe the two events in Trabuco 21 Canyon in 1975?
22 A.
The California Institute of Technology has indicated i
fj 23 that these earthquakes occurred on January 3, 1975, at 24 5:54 and 6:01 Greenwich Mean Time, were located several 25 kilometers west of the Christiantitos fault, and had k'n s 26 local magnitudes of 3.8 and 3.3 respectively.
Portions NY 1
4
'O
O 1
of Mission Viejo, Laguna Niguel, and San Juan Capistrano c(3
)
f~
2 reported the occurrences; howe 9er, no damaga was U
3 reported.
The strong motion instruments at San onofre 4
Nuclear Generating Station, approximately 20 kilometers 5
away, were not triggered indicating that ground motion
,u 6
had attenuated to less than 0.01g.
A field survey in a 7
region around the epicenters and along the Cristianitos 8
fault did not locate any ground surface rupture.
g 9
Q.
Were you requested by Applicants to perform a special 10 study to refine the earthquake locations?
11 A.
Yes.
ov 12 Q.
Did you prepare a repo -t of your study?
13 A.
Yes, I prepared Exhibit SB-1; " Seismological 14 ggg Investigations of the San Juan Capistrano Area, Orange 15 County, California, July 1975."
16 Q.
Would you describe the results of that study?
17 A.
Yes.
The crustal velocity model used to locate O
l 18 earthquakes by Caltech is a regional average that is 1
19 more or less suitable for all of southern California.
I 20 g
developed a refined crustal model for a more limited 21 region to more accurately locate the epicenters and to 22 fix limits on the hypo entral depths.
This new velocity 23 model was developed from data obtained from two g
l 1
24 calibration blasts that were located slightly east of 25 the two epicenters.
Although the blasts were fairly i
l 26 small, 700 and 1400 lbs. each, clear compressional wave 5
10
O rw I
arrivals were obtained at seismic stations located at 8_,
2 distances of over 120 kilometers from the blasts.
)
3 Altogether 12 re Jional stations recorded both the blasts 4
and the small earthquakes.
The improved locations are 5
g shown on Figure 16 of SB-1 (copy attached).
Epicentral 6
location error is in general a difficult parameter to 7
estimate accurately.
HYPO 71, a computer program g
8 developed by the U.S. Geological Survey and used to 9
locate these events, estimates a horizontal uncertainty 10 of 0.7 km and 1.2 km for the two events.
In view of the o
11 number of stations used in the locations, and the high 12 quality of the velocity model, this epicentra: error 13 estimate is reasonable.
Depth estimates for both events (l) 14 range from 2.0 km to 4.6 km.
HYPO 71 depth error 15 estimates for both events are 1.3 km and 2.6 km.
16 Q.
Did your studies address the possibility of spatial O
17 association of these events with the Cristianitos fault?
18 A.
Yes.
19 Q.
Would you describe the results of that work?
O 20 A.
These events cannot be spatially associated with the 21 Cristianitos fault.
Figure 19 of SB-1 (copy attached) 22 shows a geologic cross-section through this area.
This O
23 figure snows the location of the Cristianitos fault, the 24 available well logs showing the geologic sections and 25 the deepest hypocentral determinations for the two G
26 events.
Assuming the shallowest possible dip for the 6
'O
O 1
Cristianitos fault, defined by the location of the 7
(
/
2 surface trace and passing just beneath the logged wells, 3
ths.'e events could not lie on the Cristianitos fault 4
plane.
5 Q.
Although your testimony is that these two events are not 6
spatially associated with the Cristianitos fault, is 7
there any evidence that suggests these events share a 8
style of faulting cimilar to the Cristianitos fault?
yy 9
A.
No.
First motion readings from 30 stations surrounding 10 the two events have been used to compute focal 11 me hanisms.
The focal mechanism solutions and the dat'.
O 12 from which they are derived are presented in Figures A6 13 and A7 (copies attached) in SB-1.
These two events are 14 ggg strike-slip with a significant thrust component.
Motion 15 on either of the two acceptable fault planes is 16 estimated to rake 40 The motion on either plane is 17 oblique to the trend ;f the Cristianitos fault.
The nv 18 relative orientation of the two earthquakes, their 19 closeness in time, and tpace, and the identical focal 20 O
mechanisms strongly suggest the northeast trending plane 21 is the actual fault plane.
This plane is approximately 22 parallel to Trabuco Canyon and dips 53 northwest.
23 g
Motion on this plane is thrust left lateral, oblique 24 slip.
These two events also lie along the trend of 25 Trabuco Canyon, a significant geomorphological feature.
26 The focal mechanism indicates that the motion is not the lg_v) i I
l l
7 O
l
O 1
same as would be expected from the Cristianitos which cs
(
)
O,'
2 was dip-slip.
The spatial separation of these two 3
events from the Cristianitos fault and the reversed 4
sense of motion strongly support the conclusion that 5
these events were not associated with the Cristianitos g
6 fault.
7 Q.
You earlier indicatad that a purpose of your testimony 8
is to assess the significance of the cluster of g
9 microseismic events that occurred in Trabuco Canyon in 10 1977.
" lease describe your studies of those events and 11 their results.
g 12 A.
A small sequence of five events was detected by the 13 CalTech seismic array in 1977.
The largest event was 14 gggp magnitude 2.8.
The cluster started on June 29 and the 15 last event was recorded on July 1.
The relative arrival 16 times of the compressional waves at the various stations 17 of the Caltech array suggest that these events occurred g
18 within a volume with dimensions of only a few hundred 19 meters.
I have evaluated these events using the 20 velocity model developed for evaluation of the January O
21 1975 events.
The average location is about 2.5 km north 22 of the 1975 events and is acain within Trabuco Canyon.
O 23 Because the events are so small, there is insufficient 24 first-motion data to reliably constrain the focal 25 mechanism.
Although the strike or the focal planes is (1
26
///
3 c
1 xs 8
O
O 1
unknown, the data are consistent with a thrust mechanism 2
which is consistent with the 1975 events.
O 3
Q.
What conclusions have you reached from these 4
investigations?
5 A.
The data from the 1977 cluster and the 1975 events O
6 strongly support association of these events with each 7
other and in turn with the alignment of the 8
O northeasterly trending Trabuco canyon and not the 9
Cristianitos fault.
10 11 O
12 13 g
14 15 16 O
17 18 19
'O 20 21 22 f0 23 l
24 25 26 9
(O
(
l
G Attr, chm:nt 1
DR. SHA'.IN BEIHLER
(,,)
2 PUBLICATIONS o
3 Refereed Papers 4
1.
Press, Frank and Biehler, Shawn, 1964.
Inferences on 5
crustal velocities and densities from P wave delaye
,V 6
and gravity anomalies.
Journal of Geophysical 7
Research, v.
69, no. 14, pp. 2979-2995.
8 2.
Biehler, Shawn, Kovach, R.
L.,
and Allen, C.
R.,
1964.
9 Geophysical framework of northern end of Gulf of 10 California structural province.
American 11 Association of Petroleum Geologists Memoir, pp.
g 12 126-143.
13 3.
Biehler, Shawn, 1965.
Gravity studies of Southern 14 California.
Transactions American Geophysical qggg 15 Union, v.
46, no.
- 3. pp. 557-550.
16 4.
Biehler, Shawn and Bonini, W.
E.,
1969.
A regional 17 73 gravity study of the Boulder Batholith, Montana.
18 Geological Society of America Memoir 115, pp.
19 401-422.
t 20 5.
Solomon, Sean and Biehler, Shawn, 1969.
Crustal O
21 structure from gravity anomalies in the southwest i
12 Pacific.
Journal of Geophye.ical Research, vol. 74, 23 no. 27, pp. 6696-6701.
g 24 6.
- Sleep, N.
S.
and Biehler, Shawn, 1970.
Topography ano i
25 tectonics at the intersection of fractu re zones l
26
///
lg$
l
{
10 O
O 1
Refereed Papers (Continued) g-2 t;ith central rifts.
Journal of Geophysical 3
Research, vol. 75, no. 14, pp. 2748-2751.
4 7.
Biehler, Shawn and Rex, R. W.,
1971.
Structural geology 5
and tectonics of the Salton trough, Soutnern 6
California, pp. 30-42 in Geological Excursions in 7
Southern California.
University of California, 8
Riverside, Campus Museum Contributions #1.
.g 9
8.
- Elders, W.
A.,
- Rex, R. W.,
Meidav, Tavi, Robinson, P.
T.,
10 and Biehler, Shawn, 1972.
Crustal spreading in 11 S uthern California.
Science, vol. 178, pp. 15-24.
0 12 9.
- Hanna, W.
F.,
- Rietmann, J.
D.,
and Biehler, Shawn, 13 1975.
Los Angeles Sheet.
Bouguer Gravity Map of 14 California, California Division of Mines and ggg 15 Geology.
16 10.
- oliver, H. W.,
- Robbins, S.
L.,
- Grannell, R.
B.,
- Alewine, 17 R.
W.,
and Biehler, Shawn, 1975.
Surface and g
18 subsurface movements determined by measuring 19 gravity, Ch. 16 in Oakeshott, G.
B.
(Ed), The San g
20 Fernando Earthquake of February 9, 1971.
21 California Division of Mines and Geology Bulletin, 22 196, pp. 195-212.
O 23 11.
- Elders, W.
A.
and Biehler, Shawn, 1975.
Gulf of 24 California rift system and its implications for the 25 tectonics of western North America:
Penrose 26 Conference Report.
Geology, vol.
3, pp. 85-87.
11 0
O 1
Refereed Papers (Continued)
I )
2 12.
Rotstein, Yair, Combs, Jim and Biehler, Shawn, 1976.
O 3
Gravity investigation in the southeastern Mojave 4
Desert, California, Geological Society of America 5
Bulletin, vol. 87, pp. 981-993.
6 13.
- Biehler, S.
and T-C Lee, 1979.
Resource Assessment in 7
Geothermal Energy and Regional Developmsnt, edited 8
by S.
Edmounds and A.
Rose, Praeger Scientific.
, _,v 9
14.
- Biehler, S.
and Rotstein, Y.,
1979.
Salton Sea Sheet, 10 Bouguer Gravity Map of California.
California 11 Division of Mines and Geology.
O 12 15.
- Crowell, J.
C.,
- Beyer, L.
A.,
- Biehler, S.,
- Ehlig, P.
L.,
13
- Hall, E.
A.,
- Junger, A.,
and Vedder, J.
G.,
1980.
14 Geologic cross sections from Patton Ridge to the gg 15 Mojave Desert, across the Los Angeles Basin, 16 Southern California.
Geological Society of 17 America, Map and Chart Series MC-28K.
O 18 19 20 O
21 22 23 0
24 25 26 C\\.
(_)
12 O
g 1
Abstracts
'i 2
A-1 Biehler, Shawn.
Geophysical investigations in the O
3 Salton trough, Southern California. Trans.
A.G.U.,
4 vol. 44, no.
1, 1963.
5 A-2 Biehler, Shawn and Press, Frank.
P wave anomalies in O
6 California as an indication of crustal structure.
7 G.S.A.,
Sp. paper 72, 1963.
8 A-3
- Press, F.
and Biehler, Shawn.
Velocity reversals in O_
9 California batholiths.
Trans.
A.G.U.,
vol. 45, 10 no.
1, 1964.
11 A-4 Biehler, Shawn.
Part II, Application to a gravity C
12 study of a portion on the San Jacinto fault sone.
13 Abstracted presented at S.E.G.
annual meeting.
e'g 14 A-5 Biehler, Shawn and Kovach, R.
L.
Gravity study of the x) 15 San Jacinto Valley.
Soc. Expl. Geophys. Ann Mtg.,
16 1965.
17 A-6 Press, Frank and Biehler, Shawn.
Thermal argument for 18 a velocity reversal in the Sierra Nevadan crust.
19 G.S.A.,
Sp. paper 82, 1965.
20 A-7 Biehler, Shawn and Bonini, W.
E.
A geophysical n
Q,Y 21 interpretation of the boulder batholith, Montana.
22 Trans.
A.G.U.,
vol. 47, no.
1, 1966.
23 A-8 Biehler, Shawn.
Relation of gravity anomalies to g
24 geologic structure in Southern California.
Trans 25 A.G.U.,
vol. 49, no.
4, 1968.
26
///
13 O
l
~
1 Abstracts (Continued)
)
~~'
2 A-9 Solomon, Sean and Biehler, Shawn.
Crustal structure O
3 from gravity anomalies in the southwest Pacific.
4 Trans.
A.G.U.,
vol. 50, no.
4, 1969.
5 A-10 Folinsbee, R. Allin, Biehler, Shawn, and Bowin, Carl O.
g 6
Gravity anomalies and crustal structure of 7
Trinidad.
G.S.A.,
Ann. Mtg., Atlantic City, 1969.
8 A-11 Roddy, David J.
and Biehler, Shawn.
Recent geological g
9 and geophysical studies of the Flynn Creek 10 structure, Tennessee.
G.S.A.,
Sp. paper 87.
11 A-12 Elders, Wilfred A.,
Rex, Robert W.,
Meidav, Tavi, 3
12 Robinson, Paul T.,
and Biehler, Shawn.
A plate 13 tectonic model for the Salton trough, G.S.A.,
}
14 Cordilleran Section meeting, Riverside, 15 California, 1971.
16 A-13 Biehler, Shawn.
Gravity models of the crustal 17 3
structure of the Salton trough, G.S.A.,
18 Cordilleran Section meeting, Riverside, 19 California, 1971.
3 20 A-14 Grannell, Roswitha B.
and Biehler, Shawn.
A regional 21 gravity survey of the San Gabriel mountains, 22 California, G.S.A.,
Cordilleran Section meeting, 3
23 Riverside, California, 1971.
24 A-15 Willingham, C.
Richard and Biehler, Shawn.
Basement 25 fault geometries of the San Bernardino Valley and 26 western San Gorgonio Pass area, Southern 14
)
O 1
Abstracts (Continued)
-\\
'l 2
California, G.S.A.,
Cordilleran Section. meeting, O
3 Riverside, California, 1971.
4 A-16 Biehler, Shawn and Combs, Jim.
Correlation of Gravity 5
and Geothermal Anomalies in the Imperial Valley, 6
Southern California, G.S.A.,
Cordilleran Section 7
meeting, Honolulu, Hawaii, 1972.
8 A-17
- Oliver, H. W.,
Robbinr,,
S.
L.,
- Alewing, R. W.,
e r.d 9
Biehler, Shawn.
Gravity changes associated with 10 the San Fernando California earthquake of 11 February 9,
- 1971, G.S.A.,
Cordilleran Section mU 12 meeting, Honolulu, Hawaii, 1972.
13 A-18 Biehler, Shawr..
An Audio Visual Approach ot the Earth
,y 14 Sciences for non-science majors, Amer. Geophys.
)
15 Union, Meeting, Washington, 1974.
16 A-19 Biehler, Shaan.
A Possible Gravimetric Method for the 17 Prediction of Earthquakes, Amer. Geophys. Union, O
18 Meeting, Washington, 1974.
19 A-20 Rotstein, Yair, Combs, J.
B.,
Biehler, Shawn.
20 g
Geophysical Investigation in the Eastern Mojave 21 Desert, California, Amer. Geophys. Union, Meeting, 22 Washington, 1974.
23 O
24 25 lh 15 O
O 1
DR. SHAWN BEIHLER 7,
t 2
PROJECTS O
3 1.
Biehler, Shawn and Roddy, D.
J.,
1964.
Geophysical 4
study of the Flynn Creek structure, Tennessee.
5 U.
S. Geological Survey Annual Project Report, pp.
v 6
163-180.
7 2.
Biehler, Shawn, 1971.
Gravity studies in the Imperial 8
Valley, pp. 29-41 in R.
W.
Rex (Ed), Cooperative g
9 Geothermal-Geophysical-Geochemical Investigations 10 of Geothermal Resources in the Imperial Valley of 11 Gs, California, IGPP-UCR Report.
12 3.
Biehler, Shawn and Getts, T.
R.,
1972.
Geophysical 13 studies of the lower Borrego Valley and west Mesa cs 14 areas of Southern California, in Cooperative Lj 15 Investigations of Ouothermal Resources in the 16 Imperial Valley area and their potential value for 17 O
desalting purposes, University of California 18 Report, IGPP-UCR-72-33.
19 4.
- Robbins, S.
L.,
- Grannell, R.
B..,
- Alewine, R.
L.,
20 O
Biehler, Shawn, and Oliver, H. W.,
1973.
21 Descriptions, sketch maps, and selected pictures 87 22 gravity stations reoccupied after the San Fernando 23 O
earthquake of February 9, 1971.
U.S.
Geological 24 Survey Open File Report.
25 5.
Biehler, Shawn, 1974.
Geophysical Surveys of the Vidal 26 Nuclear Generating Station - in Information 16 O
O 1
Projects (Continued)
O 2
concerning site characteristics - Vol. III -
O 3
Southern California Edison Company.
4 6.
Biehler, Shawn, 1974.
Interpretation of regional and 5
local gravity anomalies in the Statsop, Washington
)
6 area - in WPSS Nuclear Power Project No.
3.,
7 Preliminary Safety Analysis Repcrt.
Vol. II, 8
.g Woodward Lundgren & Associates.
Revised.
9 7.
Biehler, Shawn, 1974.
Geophysical Interpretation of the 10 Subsurface Structure of Yuma, Arizona - in 11 O
Ge technical Investigation Yuma Dual-Purpose 12 Nuclear Plant, for the Salt River Project, Woodward 13 McNeil and Associates.
14 8.
Biehler, Shawn, 1974.
Groundwater Investigation of the
.c }
15 Quatal Canyon Area, California - Report prepared 16 for Kohner & Associates.
- g 17 9.
Biehler, Shawn, 1975.
Gravity Study of the Colorado 18 River Indial Reservation - Report prepared for 19 Southern California Edison Company.
g 20 10.
Biehler, Shawn, 1975.
Regional Gravity and Magnetic 21 Analysis of the Marysville area, California -
22
-Report prepared for Pacific Gas & Electric Company
,O 23 by Woodward Lundgren & Associates.
24 11.
Biehler, Shawn, 1976.
Seismological Investigations of 25 the San Juan Capistrano area, Orange County,
.g.
26 California in Recent Geotechnical Studies Southern 0
17 O
O 1
Projects (Continued
\\
)
2 Orange County, California - Southern California
.O 3
Edison Company.
4 12.
Biehler, Shawn, 1976.
A Geophysical Analysis of the 5
Proposed Sundesert Site, Palos Verde County, in
,U, 6
Sundesert Nuclear Power Plant - Early Site Review 7
Rcport, Fugro, Inc.
8 13.
Biehler, Shawn and Lee, T-C.,
1977.
A Resource 9
Assessment of the Imperial Valley - Dry Lands 10 Research Institute Report DLRI #10.
11 14.
Bi hier, Shawn, 1977.
Regional and Residual Gravity O
12 Anomalies of the Northwestern United States and 13 British Columbia - Weston Geophysical Corp.
14 15.
Biehler, Shawn, 1977.
Regional and Residual Gravity g
15 Anomalies of the Northeastern United States and 16 Parts of Canada - Weston Geophysical Corp.
17 16.
Biehler, Shawn and Roche, Steve, 1977.
A Detailed O
18 Gravity Study of the Elsinore Basin, California -
19 Report preapred for the Eastern Municipal Water 20 O
District by Miller, Harding and Lawson, Inc.
21 17.
Biehler, Shawn and Lee, T-C.,
1978.
Geothermal 22 Resource Assessment of the Imperial Valley - in 23 O
Imperial County, California - Geothermal Element.
24 18.
Biehler, Shawn, 1978.
The Subsurface Structure of the 25 Soda Dry Lake, San Bernardino County, California -
26 Southern California Edi
>n Co.
.g'i
')
s_/
18 O
O 1
Projects (Continued) g3V 2
19.
Biehler, Shawn, 1978.
A Gravity and Magnetic Lineation O,
3 Study of the California Desert Conseration Area -
4 Report prepared for the Bureau of Land Management 5
g by General Electric Co.
6 20.
Biehler, Snawn, 1978.
A Geophysical analysis of the 7
Heber Geothermal area, Imperial Valley, 8
O California - Pacific Petroleum Company.
9 21.
Biehler, Shawn, 1979.
A Geophysical Study of the Mission 10 Creek Subbasin - Report prepared for the Desert 11 Water Agency.
O 12 22.
Biehler, Shawn and Whalen, Philip, 1980.
Geophysical 13 Investigations for the Guam Lens Study - Report qgg 14 prepared for the Guam Environmental Protection 15 Agency.
16 23.
Biehler, Shawn, 1981.
Seismic Refraction Studies in o
17 Shadow Valley - Report prepared for the Moly 18 Corporation.
19 0
20 21 22 0
23 24 25 O-3 26
()
19 O
O O
lO 3
4 5
,o 6
7 8
'o 9
10 11 0
12 13 I4
- h 15 16 17
'O 18 19 O
20 21 22
'O 23 24 25 O
26 20 lO
?
,,/
(
p... l 19.
, g:rc. ) ' n
.c.
- '. :.,. ~.,,.v,..:q)t. :,
. m.....
5, 3, y 1 77,3 4
..s v
g ~'.:. y. ),
- ~
r!
.. e f-
... ' ' - gi c
<s a
Lk.,.s-? ",. L ?.., rn.,.'S;',t c. ; H) t.:--
- r ', _ ;
J '. d....r% '%
3-@-
,. a f,
y
.. <p.'., s
\\
- .... p.;,
- -c....7 9., K ~.., %.,.y,.m'g'.. R.e;, 3,.jou)k,,'o ay e
.. s.,,,
!y.
. :g.3...%-~,>.. r~:&..:. ~......,. u, - - ---, -
, _.. - e m a. 4:.: N. t
-w
- n 1
r
,,-::- p
,e
.w.
- f.. ) L \\.e).. b' r l_:.
. o
,.; s
'."?:n'
..:;..:,T.dT,-'-*/ e, %....:.Q.. - Q~r.;. [!,
\\k__a.)..
, ::> ;..k~...;;.
^~'
s
. ~~' m
....c,.
u.
..\\
,a c
,. m t.h ;e
,..?.,.
.**,f-A:. g -
,it
'-s-e, n
- }...
s f
y
, ~ _
,2 t
.).
,.6 s
. a /~~ l l c '. '. I t'
'\\
W
- % 'f % Q' z'? W M.T..S$.....,.,,.f. \\ -f..,d 7 %, e b:.l.#
~..Mr w %..s.:......
- .. n,/~'\\y-Q~.e
. h. n
,m. - L,g c;;
+ :-, W q.
.v.
,s s. '..
-, t
-./,
, g f n ?. ~, w *S R - Q,/e -..
s
-J g ) u, p1 j-
.w Q
- .5p-
'Q g #K l' ;.'%;,,,J W W ' '1 &% ="f
~"'",; W@c @
.B t: 2 r:A:$.t.....K,., : M N=, Wj.
~' s
- w..y
- t..l;~~/h.
M R,,
.-- i. e,, s9 o..
., A..
. - w...
'.,.m ;"
t w
c, s
Y
.l
..'"./
- }' ~, $ -
v-L'l C.3
.f
~
-- b -
C;;.
m.f - ~ ~%5'""}..
. _; ;l)ljf,,:(
l
- QQ.
b:
\\
't l
"'T
'*l'
, / ?"r
- C8.
g'."*,{
~.y~W..?;)'
s l
_ t-i. J. ^IQ, L.-
&yf gt ?e -n j_ p% n n ?=255 4,?% W. ~ &.
= - < Wr inir^.
4 t
'. W(l/g 5"
u_
=
hc l ri
- i..
N
? 2.5W-$$$$m&x.:...;..eg -s.3.y$. v&
$KN
-9 n..
.t bh:'WtN.
h.s.i &r.~ p a.s.,s % %p., -U..=..: &y * % d
~
OM 6.
.e eh m
~, h
%,. d.,;.p.s.,v. z v n.. -s:.-T._, m p-fQ A 7s@a:e,e,.-
-tu t; : c.
.]o.:
/~~.-. y., Q ; p r'.-:.',C..i
.p(, y't.
a f
n;c.
'i.
... e : %. f :.....%..a~h g,Nff%'N N f/$
h'k' 4
= c.-,
.. q > x ~ g v sp 8 8[u.-
I
'~
s.
o
- y s.y:33'f,Ilr'-W
_.jf.
- t; e,:
,.. r 3 y
Q. ; :.;$,,,g). v.[. mW.ml'
,.4
&wpQ Q? {.[& ].M (q
'd 'f' g
.' f,
-t.t,.,f i omav
- s. s. s, -;>
m.. ~w g f W e __ 4...j. A c w; - y. 7. w..n'-c 4, ye
'Qb. g y..x.. s %,
b, aid.
- y
.i iy
,, x
~
s.,,,""', d)/.' @ - Q r, '.'i 1g p
s s......
o
..j]'A.?s.%s in
- i-V
. f..' s.. e ' O. ? ': - -
C' N.i.r,.'.l'd. \\
.o i
s rN u
i 4.,M. a % ;w e.>,'y,f+..v%,-.3
.;3;.c. m ( h. W.= 3 S g
,e u
s-
- c. - 3 _. ;..3 s
.y.s
- - s..>u.s
- ,P s... :.:.
s -
s.
m s.
. Q.n,v p W-,s Yi----4i$D. k $.y.g ;...., i.)y - s.
3.,
l e,
,_' Q,.
>. m
- -w,._
- 3+.g o-tv -
.[
h r,?g. :
Q9 4
~
jf.i,.,1
\\r Y, ?,. --
9'. ' E
' # "[
P r~o c-- X,.i ? % ".'.
8
. s _~_ _,f
'/-
m > :. 4_m - _.
a~e s
o
- W'"* (_ %.-. c_..r..,~c..,^ W n,
-n e
g
\\
,C.'
. _ _3.g.K"*"f.5"""~{'
gg
- Q Q. f..h Q...*...?.llf-
~.
~
d,
[^f' {
,. - ~
- m. i 2/~
o-s. :-------.u s_s,m"_ "_ '4g
. el
.w s.
..- _.-=-~
,..)
r ~
__..,.fs..
- J a['3.c...'f* q'::l,g l,%\\..-w -
s,
'&.,. 5,
~ ~ ~
- i r
s
.j/ %
'd W
J 7.
.m{.)lif. M's,W C 'iM 'B'i.r'O::s Cl! 't P'-' MzN :
n
%)
SB Figure 16
{
M it!8!4
$sitt tit *2 rb o
h estensso
- h. m s ses
+
ini
~, -
7 D
3:M
~
6:n
- gr=._.
sary
=
a e
g-w l
L Capelf tasso oO Y
3 q.
~-1
~.r**w
- NI, em,
,--% O~'
f
?
36OreitEt r N
I aueN! mrOcfNi
(
N08tf teECCAA l
, Qa -
, Q~
~,
4 gowys sioctut g'
k TO pan oa g
--h..'._--...........---
' p p.
'm' f
k fKiAt?
'p waoae:S Og tGOCf NE N
4 W-
/
S000 p#
StsPt
~~ y #
=~~~~~ -
((
toctu?
4 2000-o S AN Tia G O f
f rdHJ ratractnt 4
Sllyt t ADO N-
- 4 3000-wettiaws
.gggg aw
.w CR ETAClots 6-w I
4 Geological Cross -Section After West, 1975 x
Cross - Section A-8 w
w a
a w
w fHypocenfers i
4000-e i
m.
Q
(
trrz 'nsa.1 L 843 uno. w.,a e i e,uir. u e si wu :7 eas m.,,., r....
e m m,, u
'sp
.m
... w
+
ea -.
4 4
3
' se
+
w
,. u e
sea levei ascw ?ge:,
f{ _,e
' w j y$f 5:'%'%'" ***ec,a&Q>= A :
/
l/
~J I
~_iF"
- ]ca.,%,,
//
, vim V
~$, Jj
'jfff V
i *.
se s.,
f
~
L>;
~~
I ', -
,A J !
/
l.:
d
~
p 1
/
- s
/
l u<.
I
/
j M8tDAG""
[
e W*
l-l'
/
O.
i
.f <'
/
@Ei.,
I e'
/~
'co' I
Ti
/ /) /
/
1
[
/y 3%_
),-
h' ~
/
/
/
k..'
m e s s o.,
j y
.Ly'
/
/
y
[
I l ' Y:..
, / y/
l A,,
s,
,g
, [,,
w,ma s r
/
h,l ll j
/
l
/
j
,~
i l
I 9
e-i I
l Sh oUowest possible I
l projection of j
- Cristi nitos Foult J'
l' I
I t
SB FIGURE 19 l
i i
6 l
!l2-l y
l
o 20 g.
w EVENTe 5514 1/ 3/75 o
N nf-$
Q SBB
~r m
D[ M 8D I "E * 'C GSC Q%
O
- C HI O Q$$
S 2
A "Kn taDUM
~*
-E C
S DRM O
g bn O SCI O
O CPE O
I S
SB FIGURE A6.
Event 554 - Focal
! O.
Mechanism Relative
(
to Shot 2, Model 3 0
O IO 9
r EVENTc E1 1/ 3/75 o
N
]
DTWL
,o C HIo P
Oss IRN C
O c =" "
w-
-E j
l C
S C P'"
o
] IKP g TI g mR
,\\
D "E o
O 1
S SE FIGURE A7.
Event 601 - Focal
.fechanism Relative
'to Shot 2,'!odel 3 lO
O 1
TESTIMONY OF DR.
I. M.
IDRISS
[_s\\
s"/
2 Q.
Would you please state your name?
O 3
A.
Dr.
I.
M.
Idriss.
4 Q.
Are you the same Dr.
I.
M.
Idriss who appeared as a 5
witness earlier in this proceeding?
6 A.
Yes.
7 Q.
What is the purpose cf your testimony in this portion of 8
the proceeding?
9 A.
Another issue in this proceeding is whether as a result 10 of data gathered from earthquakes which occurred 11 subsequent to the issuance of the construction permit, C,
12 the seismic design bases for SONGS 2 & 3 is inadequate 13 to protect the public health and safety.
My testimony 14 ggg in this portion of the proceeding discusses the data 15 gathered from the 1979 Imperial Valley earthquake.
16 Q.
Have you assessed the impact of the data from the 1979 17 O
Imp rial Valley earthquake on the adequacy of the SONGS 18 2 & 3 vertical DBE spectrum?
19 A.
Yes.
The impact of the data from the IV 79 earthquake 20 g
on the adequacy on the SCNGS 2 & 3 vertical DBE spectrum 21 has been assessed as detailed in Exhibit IMI-9, "NRC 22 Staff Question and Response 361.44."
As described in 23 g
this exhibit, the instrumental peaks for the IV 79 24 vertical data in the distance range of interest exceed 25 the vertical DBE in the narrow period range of 0.05 to 26 0.12 seconds by less than about 20 percent.
The amount g
a
O 1
that the instrumental spectrum exceeds the design 7s'a\\
2 spectrum is considered not to be significant for nu 3
design.
Reductions in instrumental spectra larger than 4
20% can be justified to reduce the instrumental spectra 5
g to design level spectra.
6 Furthermore, as described in Dr. Frazier's 7
Testimet.
high vertical accelerations recorded in the 8
g near field during the IV 79 earthquake were due to 9
unique subsurface conditions which caused the P-wave 10 energy to dominate the vertical component of the ground O
11 m tion.
12 In addition, because the P-wave energy arrives 13 first, there is a significant phasing difference between
-ggg 14 the peak vertical and horizontal components of ground 15 motion in the near field as described in Exhibit IMI-9.
16 Specifically, consideration of the phasing led to the g
17 development of the mean and 84th percentile vertical 18 response spectra for IV 79 at a distance of 8 km shown 19 in Figure IMI-E, " Comparison of Vertical DBE Spectrum a
20 with 1979 Imperial Valley Vertical Spectral Values 21 occurring during the Significant Duration of Horizontal 22 Shaking."
Also shown in this figure is the SONGS 2 & 3 0
23 verti cal DBE spectrum.
As can be seen in this figure 24 the SONGS 2 & 3 design basis envelops the IV 79 mean and 25 84th percentile spectra in the entire period range.
26
///
2 0
O 1
Therefore, it is concluded that in SONGS 2 & 3 0
2 vertical DBE spectrum is conservative with respect to O
3 vertical ground motions recorded at C mparable distance 4
f 8 km during the 1979 Imperial Valley earthquske, 5
O 6
7 8
O 9
10 11 0
12 13 14 15 16 17 0
18 19 20
- O 21 22 23
- O 24 25 26 3
O
. I i
.O 1000 i
i i
i 7.s.
i i
1 1979 IMPERI AL VALLEY O
M = 6.9, ML = 6.6 s
300 SONGS 2 & 3 (Vertical) -
Vertical Component Dist n e: 6 to 13 km O
100
/\\
~
J \\
^
g n'
I
\\
_j 30 84th Percentile
\\
j
/
4
\\
E
^
\\
/ \\,e/
g v
pn V W A V /^\\^\\j
\\
8
\\
.I O
y I
$ 10
/,
\\_
\\-
e
/
3-Mean
.5 I
g O
1 cE I
Y
/
O
/
O 1
~
,/
~
/
-,/
-/(
Damping = 0.02 I
'!'I I
't
I
'I 0.1 O
0.01 0.03 0.1 0.3 1
3 10 Period (seconds)
Figure IMI-E - Comparison of Vertical DBE Spectrum with 1979 Imperial Valley Vertical Spectral Values Occurring during the Significant Duration of Horizontal Shaking
.O
O 1
TESTIMONY OF DR. ROBERT L. McNEILL
( 'I 2
Q.
What is your name?
O 3
A.
Dr. Robert L. McNeill 4
Q.
Are your the same Dr. Robert L. McNeill who appeared as 5
a witness earlier in this proceeding?
6 A.
Yes.
7 Q.
What is the purpose of your testimony in this portion of 8
the proceeding?
G 9
A.
Another issue in this proceeding is whether as a result 10 of data gathered from earthquakes which occurred 11 subsequent to issuance of the construction permit, the v
12 seismic design basis for SONGS 2 & 3 is inadequate to 13 protect the public health and safety.
My testimony r]
14 demonstrates that when compared against such subsequent x;
15 earthquakes the DBE for SONGS 2 & 3 is conservative.
16 Q.
Have you examined instrumental recordings of recent 17 earthquakes to evaluate the conservatisms of the DBE?
O 18 A.
Yes.
I have reviewed every large recent earthquake for 19 which I have data, and have compared them to the DBE, as 20 I will w discuss.
O 21 The most relevant comparisons to the DEE would be 22 from several heavily instrumented, strike-slip 23 O
earthquakes of approximately M7.
On that basis, a mean 24 attenuation relationship for this kind of event could be 25 established, with appropriate confidence limits.
Those 26
///
O
O 1
I-)
several heavily instrumented ideal earthquakes are not 2
available, so the data which are available rust be used O
3 with appropriate conservatism.
One good measure of the 4
conservatism of the DBE is the IV-79 event, which was 5
strike-slip and M6.9.
The comparisons between that 6
earthquake and the DBE have been given in my previous 7
testimony in Figures RLM-Q and RLM-R.
8 g
Aside from IV-79, there are a few well instrumented 9
small earthquakes, and a few single recordings of large 10 earthquakes.
None of these recordings were made for the 11 O
nditions assumed for SONGS 2 & 3 (M7, R8, strike-12 slip), and some of them depart from those conditions 13 substantially.
Notwithstanding those differences, I c;
14 hase selected from those events the exceptional v
15 recordings, applying the following selection criteria:
16 (1)
Magnitude greater than M 6.
3 17 g
(2)
Distance to fault 20 km, and known.
18 (3)
Free-field or small-building, level-ground 19 instrumental station.
20 O
(4)
At least one component exceeds the 84th 21 percentile value predicted by the TERA 22 attenuation relationships for the magnitude 23 O
and distance-to-fault of that particular 24 earthquake.
25 (5)
Earthquake occurred since CP stage (1973).
26
///
2 O
O 1
Criteria (1) and (2) were applied to restrict the events 7
i1 2
to be near the range of applicable condtions, and to O
3 avoid the considerable uncertainties associated with 4
extrapolating from low magnitudes and/or long 5
distances.
Probably the most stringent and conservative O
6 criterion is (4):
a record must be strong enough to 7
have at least one excendance of the TERA 84th percentile 8
before that record is judged to be exceptional.
9 Following the five criteria above, six earthquakes 10 surfaced as having exceptional recordings.
These are 11 shown in Table RLM-1 " Study of Exceptional Recordings",
12 where the earthquakes are listed (Cols.
1,
- 2) with their 13 magnitudes (Col.3) and faulting style (Col. 5), and the j
14 significant distance (Col. 4) for each exceptional
'N]/
15 recording.
Also given is the measured PGA for that 16 exceptional recording (Col. 6), and, for reference, the 17 comparable calculated 84th percentile PGA (Col. 7) of 18 criter.on (4) above.
In order to bring these 19 exceptional recordings under diverse conditions 20 g
approximately to the conditions assumed for SONGS 2 & 3, 21 I have multiplied the measured PGA for each event by the 22 ratio of the 84th percentile PGA calculated for M7, R8 23 O
to the 84th percentile PGA calculated for the reported 24 magnitude and distance for that event, using the TERA 25 equation.
The resulting values, thus approximately 26 scaled to M7, R8, are shown in Col.
8.
3 O
O 1
I note that none of those scaled values exceed the
,_h
<('"/
More significantly, I elso note O
3 that their average is about 40 to 50 percent of the DBA 4
ZPA.
I consider these results to be highly significant 5
in evaluating the conservatism of the DBE:
I have u
6 selected the highest documented recordings from six 7
significant large earthquakec; and, when scaled to M7, 8
R8, none of their PGAs exceed the DBE ZPA, and their V,
9 average is only about half the DBE ZPA.
10 I have also performed the same calculations 11 O
p riod-by-period for the response spectrum and have 17 averaged the results.
A few points of the individual 13 spectra lie above the DBE spectrum, but by far most lie eT 14 below.
More importantly, the average of the spectra of r%)
15 these exceptional recordings lie at all points below the 16 DBE spectrum.
17 Q.
What conclusions do you draw from these studies?
O 18 A.
I believe that one of the best measures of the 19 conservatism of the DBE is to compare it to records of 20 O
earthquakes which are similar to the conditions assumed 21 for SONGS 2 & 3.
There is only one well instrumented 22 earthquake to my knowledge that approaches those 23 g
conditions, and that is IV-79.
That earthquake is 24 compared, on the basis of the statistics of all the 25 recordings around 8km (6-13 km) to the DBE in Figures 26 RLM-Q and RLM-R.
That earthquake is also compared, on 4
O
i 0
1 the basis of its exceptional recordings from 1 to 20 km,
(_h 2
to the DBE in Table RLM-1.
Both comparisons lie below O
3 the DBE.
I therefore conclude that these two studies 4
demonstrate the conservatism of the DBE, based on 5
comparison to an earthquake with conditions similar to 6
those assumed for SONGS 2 & 3.
7 I have also studied exceptional recordings from 8
earthquakes with conditions quite dissimilar to those 9
assumed for SONGS 2 & 3.
The results show that none of 10 the PGAs exceed the DBE ZPA, and that the spectra from 11 O
the exceptional recordings lie below the DBE spectrum at 12 all points on the average, and at the majority of points 13 on an individual basis.
I therefore conclude that this
,N 14 study demonstrates the conservatism of the DBE, based on
'L) 15 comparison to earthquakes with conditions dissimilar to 16 t'
assumed for SONGS 2 & 3.
a 17 O
It is therefore my opinion that the DBE as used for 18 the design of SONGS 2 & 3 continues to be a conservative 19 representation of the motions governing the response of 20 g
structures at this site due to a very large, nearby 21 earthquake when compared to large earthquakes occurring 22 subsequent to issuance of the construction permit.
23 O
24 25 5
O
TABLE RLM-1:
.3TUDY OF EXCEPTiO*8AL RECORDINGS x
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
'J Measured TERA mad Scaled Faulting PGA PGA PGA Date Ide>.tification M
R,km Style Mi,Ri Mi, R1 M7,R8 16Sep78 Tabas, Iran,Trans 7.7 3
Thrust 0.78g 0.71g 0.54g g"
Long 7.7 3
Thrust 0.83 0.71 0.58 17May79 Ga:li, USSR; EW 7.2 5
Threst 0.74 0.60 0.60 NS 7.2 5
Thrust 0.64 0.60 0.52 150CT79 IV-79 942/230 6.9 1
Strike-S 0.45 0.74 0.30 942/140 6.9 1
Strike-S 0.72 0.74 0.48 5054/230 6.9 2
Strika-S 0.81 0.69 0.58 5054/140 6.9 2
Strike-S 0.66 0.69 0.47 958/230 6.9 4
Strike-S 0.50 0.60 0.41 958/140 6.9 4
Strike-S 0.64 0.60 0.52 955/230 6.9 4
Strike-S 0.38 0.60 0.31 955/140 6.9 4
Strike-S 0.61 0.60 0.50 5165/360 6.9 5
Strike-S 0.51 0.57 0.44 5165/270 6.9 5
Strike-S 0.37 0.57 0.32 5115/230 6.9 10
- rike-S 0.43 0.42 0.50 5115/140 6.9 10 Strike-S 0.33 0.42 0.39 5058/230 6.9 13 Strike-S 0.38 0.36 0.52 5058/140 6.9 13 Strike-S 0. 3'J 0.36 0.52 09Jun80 Victoria,BC,N15W 6.3 2
Strike-S 0.85 0.64 0.53 23Nov80 Italian; ST-NS 6.5 18*
Normal 0.24 0.22 0.53 ST-EW 6.5 18*
Normal
- 0.35 0.22 0.78
,c,U 27May81 Mammoth 99/180 6.3 10 Normal 0.33 0.32 0.50 99/90 6.3 10 Normal 0.27 0.32 0.41 3679, Long 6.3 10 Normal 0.38 0.32 0.58 Trans 6.3 10 Normal 0.17 0.32 0.26 1
3754, Long 6.3 8
Normal 0.76 0.33 0.98 1
Trans 6.3 8
Normal 0.47 0.38 0.64 Average =
0.51g Epicentral distance, used incorrectly but conservatively, for purposes of study.
- Reported dip-slip, conservatively assumed to be normal for purposes of study.
- Or strike-slip.
1-
~s