ML16341B606
| ML16341B606 | |
| Person / Time | |
|---|---|
| Site: | Diablo Canyon  |
| Issue date: | 04/24/1981 |
| From: | Rothman R Office of Nuclear Reactor Regulation |
| To: | |
| Shared Package | |
| ML16340B657 | List: |
| References | |
| ISSUANCES-OL, NUDOCS 8104290140 | |
| Download: ML16341B606 (30) | |
Text
UNITED STATES OF AMERICA NUCLEAR REGULATORY COVi )ISSION BEFORE THE ATOt11C SAFETY AND LICENSING APPEAL BOARD In the Yiatter Of:
PACIFIC GAS AND ELECTRIC COYiPANY (Diablo Canyon Nuclear Power Plant Units 1
and 2)
Docket Nos.
STATE OF flARYLAND COUN1Y OF tlONTGGtlERY AFFIDAVIT OF ROBERT L.
ROTHMAN
/(
AFFIDAVIT OF ROBERT L. ROTHNAll I, Robert L. Rothman being duly sworn, do depose and state:
I ar a seismologist employeo by the Office of Huclear Reactor Regulation, U.
S. lluclear Regulatory Commission.
A copy of my professional qualification ss bound into the transcript of the Atomic Safety and Licensing Appeal Board Hearing bound into the record following TR 536 (see:
TR 538).
The purpose of this affidavit is to respond to the Atomic Safety and Licensing Appeal Board (ASLAB) Order of April 15, 198'I.
Ny comments are directed to the seismological aspects of USGS Open File Repc rt 81-365 of Narch 1981, entitled "Peak Horizontal Acceleration and Velocity from Strong-Notion Records Including Recoros from the Imperial Valley, California, Earthquake" (OFR 81-365).
Open File Report 81-365 is the latest in a series of reports on continuinq research by U. S. Geological Survey scientists on the subject of stron" ground motion resulting from earthquakes Earlier USGS reports on this subject are Geological Survey Circular 672, "Ground Notion Values for Use in the Seismic Design of the Trans-Alaska Pipeline System",
by R. A. Page, D. N. Boore, 1. B.
- Joyner, and H, l,'. Coulter, 1972; and Geological Survey Circular 79~~, "Estimation of Ground Notion Parameters" by D. N. Boore, H,
B. Joyner, A. A. Oliver III, and R. A. Page, 1978.
An adaitional report which was an snitial effort by USGS scientists to update Circular 795 for use in a study for the Hational II Security Council is USGS Open File Report 81-115, Scenarios of Possible Earth-quakes Affecting Najor California Popuiation
- Centers, with Estimates of Intensity and Ground Shaking, Anonymous,. 1981.
The information pertaining to near field
0 s t rong ground mot i on in Geo log i cal Survey Circu1 ar 672 was cons idered during the Hosgri earthquake reanalysis of the Diablo Canyon Nuclear Power Plant (DCNPP).
Dr, Nathan tl. Newmark (1976) pointed out that the value of 0.75g effective acceleration which he developed for the DCHPP oesign spectra is not inconsistent with the values in USGS Circular 672 for near field strong 1'otions, considering a repeated acceleration peak of several times, rather than one isolated peak.
Table 1 is a copy of Table 2 from Circular 672.
Geological Survey Circular 795 was the subject of much testimony during the reopened AS's;> hearing on DCNPP in October 1980.
In fact, in its oroer reopening the hearing the Appeal Board specifically instructed the parties to compare the horizont-1 peak acceleration values recorded from the Imperial Valley earthqua'ke of October
'I5.
1979 (IV-79) w'th predictions contained in USGS Circular 795, Figures 4, 24, 47 ano 4Q.
I did thi s, as part of my testimony in the reopened hearing following TR. 536.
At this point, it is appropriate to mention the hierarchy and relative significance of U.
S. Geological Survey pub'lications.
8ulletins and Professional Papers are definitive scientific reports which are subjected to considerable review before publications.
Circulars are publications which do not have the weight of Bulletins or Professional
- Papers, but have been subjected to peer review.
Open file reports are the lowest order in the hierarchy.
They are subject to limited technical review.
They are put out to get information into the public domain.
They can be characterized as being the current status of ongoing research.
A problem encountered by the authors of both Circular 672 and Circular 795 was the lack of available near field strong ground motion data, particularly for larger magnitude earthquakes.
In late 1979 there was a considerable increase
0
in the near-field strong ground motion data as a result of the Coyote Lake earthquak='f August 6, 1979 (N =5.9,1 =5.8) and IV-79.
We pointed out in the NRC Staff Response to Joint Intervenor's Notion to Reopen (Rothman and Kuo affidavit at 2) that the strong motion data from IV-79 is under study by the seismological community and that such studies will very likely continue for several years.
Open File Report 81-365 is one of many reports> using IV-79 data, which will be published.
Presently, the U.
S. Geological Survey has under review over thirty reports which are to be published in a Professional Paper on IV-79.
Nany other seismologists are also conducting research on earthquake strong ground motion.
The results of these researchers are and will be published in the scientific literature.
The staff maintains an awareness of this ongoing research and,,
when significant results are obtained, they are factored into the licensing considerations Strong ground motion seismology is a state-of-the-art science and its results are subject to assumptions and interpretation on which there is not always a complete concensus among researchers.
Open-File Report 81-365 presents the results of a new analysis of the earthquake strong motion data set.
This analysis took advantage of the recently available data from IV-79 and the more limited data sets from other recent earthquakes, in addition to the data previously used in Circular 795, to derive new attenuation relations.
In the new analysis the assumption was made that the shape of the attenuation curve is independent of magnitude and a moment magnitude scale was used to characterize the earthquakes.
These two assumptions affect the results of the study.
I
Since a magnitude scale not used previously in the DCNPP proceedings is being introduced a brief discussion of magnitude is in order.
Magnitude is a measure or earthquake source size using instrumental recordings of ground motion.
Different magnitude scales measure different phases of motion in different frequency ranges and care must be exercised in choosing the appropriate scale for the intended purpose.
Local Magnitude (M
) is the original magnitude scale L
developed from recordings of small earthquakes (M
< 5.0) at distances between L
20 and 600 kilometers in southern California.
It is determined using the largest motion recorded on the Wood-Anderson seismograph (free period 0.8 seconds).
It is particularly sensitive to short period horizontal motion.
ML is not applicable at distances greater than 500 or 600 km and must be used with care outside of California.
Surface wave magnitude (Ms) was developed subsequently to complement M
for L
earthquakes of greater size and at different locations.
It is determined from longer period (20 second) motion.
Richter magnitude (M) as it is commonly, but often not precisely, used is equal to M
for magnitudes less than 6 and Ms for L
larger earthquakes.
The reason M
cannot be used for larger earthquakes is due to the apparent saturation of the scale at around 7 1/4.
As an example of saturation, the San Francisco earthquake of 1906 had an estimated Ms of 8 1/4 and the ML is only estimated to have been between 6 3/4 and 7 (Jennings and
- Kanamori, 1979).
ML saturates because the amplitude of the shorter period waves which determine ML do not simply increase as fault length increases.
Ms saturates at about 8 1/2. ML and Ms for crustal earthquakes saturate for the same physical reason.
As Kanamori (1978) states, "The amplitude of seismic waves represents the energy released from a volume of crustal rock whose representative dimension is com-parable to the wave length."
For large earthquakes these shorter period amplitude
measurements do not measure gross faulting characteristics.
To alleviate this limitation in the fl and Vis scales the moment magnitude (5) was developed L
to represent the energy release of earthquakes associated with very large ruptures.
The moment magnitude is calculated from the seismic moment of an earthquake which is determined from very long period seismic waves (usually greater than 100 seconds) and can be'shown to be equivalent to the product of the physical properties rigidity, fault area and fault displacement.
It is this moment magnitude that is used in OFR 81-365, Two major issues discussed during the October 1980 ASLAB hearing were the flattening of the peak horizontal acceleration attenuation curve with decreasing distance to the fault (distance saturation) and the independence of peak horizontal acceleration from magnitude in the near field (saturation with magni-tude).
I testified that the IV-79 data tends to support the theory that the peak horizontal acceleration-distance curve flattens out at short distances.
I also testified that based on the physical consideration discussed by Hanks and Johnson (1976) and the additional magnitude data aoded by 1V-79 that in the near field peak ground accelerations for larger earthquakes are probably not linearly dependent on magnitude.
Figure 1 which is a copy of Figure 3 of OFR 81-365 shows that the new analysis results in a distance-peak horizontal acceleration relation that flattens with decreasing distance.
This agrees with the NRC staff testimony at the October 1980 hearing.
In discussing the assumption, used in the analysis, that the shape of the attenuation curve is magnitude independent, the authors presented no quantitative argument.
They simply stated that they see no compelling reason a priori why this
J'
I is not appropriate and that the data are consistent with it.
haking this assumption predetermines that the acceleration keeps increasing to ever higher values with increasing magnitude.
Other studies of attenuation of acceleration with distances have been performed which permitted attenuation curves whose shape could be magnitude dependent.
Campbell (1980) performed a study of'he strong motion data set (including IV-79) in which it was not assumed that the acceleration increased with magnitude in the near field.
One of his conclusions is that the results of his study have established that accelerations tend to saturate with increasing magnitude at small distances.
Hadley and others
( 1981) have conducted numerical modelling studies to investigate peak acceleration attenuation relations.
1n their study, the shape of the acceleration attenuation was permitted to be a
function of the earthquake magnitude.
They showed that the shape of the average peak acceleration versus distance curves are well described by an equation of the form P.A.~(R +
C (t'I)) 'here R is the closest distance to the fault surface
-l. 75 and C(4.5)=6, C(5.5)=12, C(6.5)=22 and C(7.0)=36km. Figure 2 is a copy of their Figure 13 and shows the normalized attenuation curves for four magnitudes.
The trend toward flattening of the attenuation curves with increasing magnitude can be seen.
This modelling study did not incorporate any non-linear near surface
- effects, so the behavior of C (tl) cannot be attributed to near site material properties.
The authors concluded that C is related to the physical dimensions of the fault and to the increasingly
'iong period characteristics of the seismic source with increasing magnitude.
The appropriateness of using moment magnitude in an analysis of near field acceleration is open to question.
The moment magnitude scale was developed because of the saturation of the magnitude scales which are based on shorter period waves.
The saturation of the "Short Period" l"L scale and its non-linear relation to the
longer period magnitude scales above hL
= 6 (Kanamori, 1979; Kanamori and
- Regan, 1981) argue strongly against the assumption that short period accelerations, particularly those in the near field, can be assumed to scale simply with H. for magnitudes in the range of interest to the OCNPP (7.5).
If this is true, the use of moment magnitude in this context may result in the over estimation of peak acceleration in the near field for large earthquakes.
Prior to the reopened hearing on DCNPP in October 1980 the ASLAB instructed the parties to compare IV-79 peak horizontal accelerations with several acceleration attenuation predictions.
Figure 47 of U.S.G.S. Circular 795 was one of these.
Figure 3 is a copy of Figure 47 from Circular 795 on which is p'totted the 50
Exceedence curves for Magnitude 6.5 from Figure 3 of OFR 81-365.
The main source of near field data is the IV-79.
This figure shows the over estimation inherent in previous studies which did not have large earthquake near field data but were extrapolations of smaller earthquakes and from larger distances.
The magnitude 6.5 curve was chosen to make this comparison because it was the highest magn>tuae for which there was near field data.
Even with reservations as to the assumptions made in OFR 81-365 we can still use some oi its results for comparison with previous work.
With respect to DCNPP extrapolation to magnituae 7.5 at a distance of 5 kilometers is necessary.
As previously noted, Newmark said that Circular 672 was not inconsistent with his effective design acceleration.
To see whether OFR 81-365 would have any effect
on this juogement we can make a comparison between it and Circular 672.
Table 1 is a copy of Table 2 of Circular 672.
Comparison cf peak acceleration of Table 1 with curves of Figure 1
shows that the results of the new study do not in general exceed those of the earlier study.
As an example, from Figure 1 the acceleration predicted for a magnitude 7.5 earthquake at 5
kilometers at the 50 percent exceedence is 0.80 g and the S4 percent exceedence is 1.4Sg wl i le Table 1 predicts an acceleration of 1.15g in the near field for a magnitude 7.5 earthquake.
It is important to note that the authors of OFR Sl-365 are quite clear in stating that "For distance less than 40 km from earthquakes with Agreater than 6.6 the prediction equations are not constrained by data and the results should be treate'ith caution."
(see:OFR 81-365,
- p. 15.). This indicates that extraoolatinn 1
to magnitudes greater than 6.6 in the near field may result in incorect accelei a.ions.
In conclusion I would like to reiterate that:
l.
OFR 81-365 is a report of ongoing research ano its results may be subject to change under further analysis.
2.
The analysis uses some controversial assumptions such as:
A - magnitude inoependent shape for regression curves B - plotting of near field acceleration as a simpie function of moment magnitude.
3.
The new analysis does not contradict the decisions made by the staff and its consultants in the Hosgri reanalysis of the DCNPP nor affect my i.estimony provided at the October 1980 ASLAB hearing.
References
- Campbell, K. M., (1980), Attenuation of Peak Horizontal Acceleration within the Near-Source Region of tloderate to Large Earthqua'kes, lERA Technical Report 80-1, TERA Corporation,
- Berkeley, CA.
Hadley, D.ti., D.V. Helmberger, and J.
A. Orcutt,
( 1981),.Peak Acceleration Scaling Studies, Sponsored by the U. S. Geological Survey Contract No. 14-08-0001-19131.
Hanks, T.C.
and D. A. Johnson, (1976), Geophysical Assessment of Peak Accelerations, Bull. Seis.
Soc.
Am., vol 66, 959-968.
- Jennings, P.C.,
and H. Kanamori, (1979), Determinations. of local magnitude HL
, from seismoscope
- records, Bull. Seis.
Soc.
Am., vol. 69, 1267-1288.
i;anamori, H., (1978i, 8ualification of Earthquakes Nature, vol. 27, 411-414.
- Kanamori, H. (1979),
A Semi-empirical Approach to Prediction of. Long Period [lotions from Great Earthquakes, Bull. Seis.
Soc.
Am., vol. 69, 1645-1670.
- Kanamori, H.
and J.
- Regan, (1981 ), Long Period Surface Haves Generated by the Imperial Valley Earthquake of 1979.
In press, U.S.G.S.
Professional Paper on Imperial Valley Earthquake.
Newmark, N.N., (1976),
A Rationale for Development of Design Spectra for Diablo Canyon Reactor Facility, Supplement No.
5 to the Safety Evaluation of the Diablo Canyon Nuclear Power Station Units 1
ano 2, U.S.N.R.C.
'Hashington, D.C.
Tobe ~ 2.
hirer fevll hiorisc rt~ I ground molion Veloc lit fern(s ~ c I P ~ ~ I ~ h soli,'I ~
e ~ Iv ~ I Acc ~' retire IOI O ~ el ehre Vt~ vervet her 9hiIUde DiSC!~ 1 ~ ~ ~ ~ I C<< ~tiov 1
~ s ~ CI I s1 I.25 I.20 l.l5 I.OS 0.90 O.d5 Td I.I5 I.IO I.OO 0.90 0.75 0.30 (th I.OO 0.95 O.eS 0.75
- 0. 60 0.20 lOss 0.75 0.70 0.55 0.55 O.ds O. I5 lss ISn I45 l35 l20 IO1 Sp Td 130 l25 I IS IOO IU 8.5 1.0 7.5 7.0 e.s 5.5 IIO I05 IOv 65 70 30 IOO IS 70 55 g
I IS 90 60 40 75 IT ITime intervei betveen Ditsl end lest eel Not ~ I I, itch pee I ol ~ litotes ~
~ cc ~ I~refior
~ O
~ t to or O ~ ~ et ~ 1 II en 0.0I e ic serves
~ re besed or instrvrnr ~ ses dei ~
9
- l. Tne vetue ~ in this teb
~
e for sites et <<hich pro d
I' bl ~
~te for
~
1 no' hr'isr I~ I cornt oncet I I rn li
~
I
~
r ic ~
~
~ dbtenc ~ ol ~ fe>> fl I ~ Im r'ne
< ~
~ ~ ~. ~ ~ Ievrt.
he prrsenc f
irn me icr is
~ oi stre*
Is
~It~ red bi
~ ~ si ~ 'n ~ cr' r ~ sts in tn ~
~ I~ ssic pic pe'
~ s v i~ r tr ~ Ic ~ ~
oe. rr o i e
~ o I rvct r ~ s.
~ "d c
<<t ~ i~ no terror s ~ I~Iiv tc Ine
~ ~ I
~
~, ',,,,~~,"...~,.~, ~
~
~'
v' or srnt err ~ nc ~ o'"~
~"""~ t' C res'O"eo of d pie c I
f rr' "~h' 'o'" ~o" I 'o""
"~
~
0 disci ~ r enient ~ Irr '
cn
~I' ctt ~
cornt'events siss I' 1 o
~ o' es ~>> ~ en ID Iv Table 1 (After U.S.G.S. Circular 672)
50 PERCENT EXCEEDANCE 8il PERCENT EXCEEDANCE 1.o zO I-(C UJ 0.1 CJ (C
(C I
O Q.Q1 CLa 7.5 7.0 6.5 6.0 5.5 M
- 5.0 M
$ 5.0 7.5 7.0 6.5 6.0 5.5 QAx)1 10 DISTRNCE
~ 10 0 I STRNCE r xv Figure 3.
Predicted values of peak horizontal acceleration for 50 and 84 percent exceedance probability as functions of distance and moment magnitude.
Figure 1 (After USGS OFR 81-365)
- 0. 3 O. 1 M
= 7 I
6 I/2 5 I/2 I/2
- 0. 03 10 30 100 CLOSEST DISTANCE (km) 300 Fichu c
13.
Summary of normalized a'ttenuation curves from the simulation work.
F'intlro
'P (After 'Hadlev anc; other s, 1981)
z0 0.1 U
Z0 0.01 z
0.001 10 100 DlSTANCE, lN KtLOhhETERS Figvre 47.
Proposed re1ations of peak horizon-ta1 acceleration to distance from slipped fault for magnitude 6.6 earihqvake.
Curve labeled S
is given by Schnabe1 and Seed (l973) for rock sites, curve labeled D is given by Donovan
(\\973) for soil sites, and curves 1abeled TO and T2 are mean curves given by Trifunac (ig76) for soft and hard
- sites, resp(.cti ve ly.
Soli d lines shoe 70 percent prediction interval for data set for, magnet ude c 1 as s 6.0-6.4 and small atrueturee,'rOm thug repOrt,'urve 81-365 is A = 6.5, %0i 7xceedence from OI.R'1-365.
t'igure. 3 (After USGS. Ci'rcular. 395$
'I
Sl.'.scrihed and suborn to before o ert L.
othman m-tnis.-'"'ay of April, 1981 No.ary Pub l'.c hy Commission expires:
/