ML20041D756
| ML20041D756 | |
| Person / Time | |
|---|---|
| Site: | Summer  |
| Issue date: | 03/05/1982 |
| From: | Goldberg S NRC OFFICE OF THE EXECUTIVE LEGAL DIRECTOR (OELD) |
| To: | |
| References | |
| NUDOCS 8203090125 | |
| Download: ML20041D756 (92) | |
Text
o UNITED STATES OF At1 ERICA NUCLEAR REGULATORY CD1tilSSION BEFORE THE ATOMIC SAFETY AND LICENSING BOARD In the fiatter of
)
)
SOUTH CAROLINA ELECTRIC & GAS
)
Docket No. 50-3 N
COMPANY
)
)
8 g
(Virgil C. Summer ?!uclear Station,
)
"a..,sE Uh3 3
Unit 1)
)
p MilR8 Igegw :
SNfDTP~f (A
NRC STAFF PROPOSED FIflDIllGS OF FACT AND CONCLUSIONS
- 3 0F LA'l Ifl THE FORfi 0F AN INITIAL DECISION Steven C. Goldberg March 5,1982 Counsel for NRC Staff t,Tr.~nsn c"TcIriL
&_7 !
^ ' *' ' ' c 1 -
8203090125 820305 y
DR ADOCK 05000 i
. o TABLE OF CONTENTS PAGE I.
Introduction and Background...............................
1 II. Matter in Controversy.....................................
2 A.
June 1981 Hearing Session............................
2 1.
Introduction....................................
2 2.
Reservoir-Induced Seismicity....................
4 a.
Maximum Reservoir-Induced Earthquake.......
4 b.
Seismic Design.............................
10 3.
Cha rl eston Earthquake..........................
12 4.
Wa te ree C re e k Fa ul t.............................
15 5.
Continued Microseismic Monitoring...............
17 B.
Subsequent Developments..............................
18 C.
Ja nua ry 1982 Hea ring Session.........................
19 1.
Board Consultants' Reports......................
21 a.
Joyner-Fl etche r Report.....................
21 (i)
!!aximum Magnitude....................
21 (ii) Ground Motion........................
23 i
l b.
Trifunac Report............................
26 i
(i)
Probabalistic Analysis...............
26 l
(ii) Response Spectra.....................
28 l
(iii) Seismic Design.......................
29 c.
Luco Report................................
31 (i)
Ground Motion........................
31 l
(ii) Response Spectra.....................
33 (iii) Seismic Design.......................
35 l
l l
I l
l I
i l
O.
PAGE i
2.
Applicant's Affirmative Case....................
36 a.
tia ximum Ma gn i tu de..........................
36 b.
Ground Motion..............................
40 c.
Seismic Design.............................
42 d.
Applicant's Evaluation of Board Consultant Reports.........................
46 (i)
Evaluation of Joyner-Fletcher -
Report...............................
46-(ii) Evaluation of Trifunac Report........
48 (iii) Evaluation of Luco Report............
51 3.
Sta ff Affi rmative Case..........................
52 a.
Maximum Magnitude Earthquake...............
53 (i)
Normal Depth.........................
53 (ii) Shallow Depth........................
54 b.
Ground Motion Estimates M = 4.5 RIS.......
56 L
(i)
Hanks-McGuire Mode1..................
56 (ii) Response Spectra.....................
60 (iii) Mammoth Lakes Site-Specific Data.....
60 (iv) Peak Acceleration Estimates for M = 4.5 Earthquake..................
61 L
c.
Ground Motion Estimates: Shallow RIS......
61 d.
Status of October 20, 1981 Board Notification...............................
65 i)
Acceleroneter Records................
65 ii)
Field Tests..........................
66 e.
Se i snii c De s i g n.............................
68 f.
Staff Response to Board Consultants' Reports....................................
73 (i)
Staff Evaluation: Joyner-Fletcher Report...............................
73 (ii) Staff Evaluation: Trifunac Report...
77 (iii)StaffEvaluation: Luco Report.......
78
r a
--111 I
PAGE F
9,. -
Additional Comments of Dr. Carl A. Newton..
79 h.
Coments of Dr.. Andrew Murphy..............
81 f
III. Principal Findings.....t...................................
82 IV. Conclusions of Law........................................
85 V.
Order...................................,,,,,,,,,,,,,,,,,,
g5 3
\\.
i
, '/ #,
)
).
f b
c.
}
1 r
s
./,
+
i i
r y'# '.
Y r
I a
v i
i J
l I
'I
-- -~--
a
.v..e,,n..,,, -. ~,. -, - - - - -
..~.,---...-.-----...,-.-,-------..--,....n.
UNITED STATES OF AMERICA NUCf. EAR REGULATORY COMMISSION BEFORE THE ATOMIC SAFETY AND LICENSING BOARD In the Matter of SOUTH CAROLINA ELECTRIC & GAS Docket No. 50-395 COMPANY (Virgil C. Sunner Nuclear Station, Unit 1)
NRC STAFF PROPOSED FINDINGS OF FACT AND CONCLUSIONS OF LAtt IN THE FORM OF AN INITIAL DECISION The NRC Staff, in accordance with 10 CFR 62.754, proposes the foliowing supplemental findinns of fact and conclusions of law.
I.
INTRODUCTION AND BACKGROUND 1.
Evidentiary hearings were held in the captioned proceeding on June 22-26, June 29-July 3, July 13-17, September 22-24, 1981 and January 11-16 and 20,1982. The record was closed on three of the five contentions at issue following the July 13-17, 1981 hearing session. The record remained open on Contentions 4 (seismicity) and 8 (emergency planning). Following the September 22-24, 1981 session, the record was closed on contention 8.
Proposed findings on contention 8 were submitted on October 26, November 23 (Applicant), November 3 (Intervenor) and November 13(Staff). The record was reopened on January 5, 1982 to receive additional evidence on a limited aspect of contention 8 pursuant to the Intervenor's motion of December 8,1981. This led to a January 20, 1982 hearing session after which the record was again closed.
A further hearing session on Contention 4 was held from January 11-16,
O
. 1982. The record on that issue was closed thereafter. Proposed findings on this contention are given below.
II. MATTER IN CONTROVERSY Contention 4: Seismicity Contention 4 states: The FSAR is inadequate with respect to the description of seismic activity in the area of the plant site. Site seismicity monitoring should be continued through 1983.
A.
June 1981 Hearing Session 1.
Introduction 2.
At the constructinn permit stage, values for a safe shutdown earthquake (SSE)I/ of 0.15g horizontal ground acceleration for structures founded on rock and 0.25g for soil foundations and an operating basis earthquake (OBE)2/ of 0.10g for rock and 0.15g for soil foundations were established. These values were arrived from the ground motion of the design basis tectonic earthquake. Tr. 767, 833.
Initial evidentiary consideration of Contention 4 encompassed a thorough evaluation of regional and site geology and seismology. This consideration centered around the following main issues:
(1) the significance of microseismic activity associated with the 1977 impoundment of the Monticello Reservoir upon the
-1/
The " safe shutdown earthquake" is that earthquake which produces the maximum vibratory ground motion for which certain structures, systems and components are designed to remain functional.
10 CFR Part 100, Appendix A, QIII(c).
2_/
The " operating basis earthquake" is that earthquake which produces the vibratory ground motion for which those features of a nuclear power plant necessary for continued operation without undue risk to the health and safety of the public are designed to remain functional.
10 CFR Part 100, Appendix A, $I11(d).
3-seismic design basis for the plant; (2) recent hypothesis and geologic and seismic findings relating to the causes of the Charleston earthquake of 1886; (3) the possible association of a projection of the newly mapped Wateree Creek fault and reservoir induced seismicity; and (4) the desirability of continued monitoring of microseismic activity in the vicinity of the Monticello Reservoir.
3.
The Applicant's initial direct case on this contention consisted primarily of the prefiled testimony of Dr. Shelton S. Alexander (Pennsyl-vania State University) (following Tr. 728), Dr. Robin K. McGuire, (Ertec) (following Tr. 730), Dr. Chang Chen (Gilbert / Commonwealth)
(following Tr. 733) and Applicant Exhibits 1-4.
This testimony was supplemented by the additional appearance of Dr. Pradeep Talwani (University of South Carolina), and Messrs. James G. McWhorter and William G. Smith (Dames and Moore) who responded to questioning by the Board and parties on related matters.
4 The Staff's initial direct case on this contention consisted of the Staff Safety Evaluation Report (Staff exhibit 1) Sections 2.5.1-2.5.3, 3.7, 3.8, 3.10; Appendices D and E, and supplement thereto (Staff Exhibit la) Sections 3.7, 3.10, 18.0 (Item 1) and Appendix D, plus Staff Exhibit 2, an internal NRC memorandum from Dr. A. Murphy to Dr. R.
Jackson, dated February 6, 1981. Appendix D to Staff Exhibit la consisted of a letter report from Mr. J. C. Mark, ACRS, to then Chairman Hendrie, dated March 18, 1981. These exhibits were sponsored by a panel consisting of the following individuals with parenthetical references to the summaries of their testimny: Dr. Phyllis Sobel (Tr. 1167-69),
. fir. Richard McMullen (Tr. 1062-63), Dr. Andrew Murphy (Tr. 1063-65),
fir. James Knight (Tr. 1067-70) (NRC), Dr. Carl A. Newton (Los Alamos Scientific Laboratory) (Tr. 1065-67) and fir. Robert H. florris and Dr. Gregory S. Gohn (USGS) (Tr. 1070-71).
5.
The Intervenor presented no direct evidence on this contention.
The Board also had the benefit of several written reports by ACRS consultants conducting its examination of Applicant and Staff witnesses.
2.
Reservoir-Induced Seismicity a.
Maximum Reservnir-Induced Earthquake 6.
The Summer site is adjacent to the Monticello Reservoir, created as part of a planned electric power generating complex. The lionticello Reservoir stores water for a pumped storage facility, provides cooling water for the nuclear plant, and serves as a makeup source for emergency cooling water. Prior to filling the reservoir, the USGS Seismograph Station at Jenkinsville (3 miles east-southeast of the site) had recorded about one local event every 6 days from 1973 to 1977.
In December of 1977, a four-station seismic network was installed in the area of the Monticello Reservoir by the Applicant. Teledyne Corporation was contracted to analyze the data. Quarterly reports of seismic activity are submitted to the NRC. Filling of the reservoir began on December 3,1977 and was completed on February 8,1978. A strong motion accelerometer was installed by the USGS in February,1978 on an abutment of Fairfield Dam.
In May 1978, the USGS began a six station seismic monitoring network in the area.
In January,1979, Dr. Talwani took over the contract to analyze the Applicant's seismic data. Dr. Talwani is
. also contracted by the USGS to analyze seisnic data from the USGS network near Monticello Reservoir. Staff Exhibit I at 2-22.
7.
An increase in seismicity near Monticello Reservoir began during the last week of December 1977. The peak activity occurred in February and liarch 1978 after the completion of filling.
In general, seismic activity has decreased since March 1978, interrupted by three swarm episodes during August to December of 1978, October 1979 and July to August 1980. As of June,1981, the maximum magnitude earthquake associated with the filling of the reservoir was the August 27, 1978 local magnitude (11 ) 2.8 event.
Staff Exhibit I at 2-21 to 22.
L 8.
The Applicant took the position that the upper bound for reservoir-induced seismicity (RIS) at Monticello is an event with Mg equal to 4.0.
The basis for this position is surriarized as follows in the prefiled testimony of Dr. Alexander:
(1) the seismicity induced by lionticello Reservoir is shallow (less than 3 km) and closely associated with the peripheries of shallow, plutonic rock bodies of limited size (approximately 1-2 km), where there is a highly variable, heterogeneous stress field and heterogeneous rock properties, both of which limit potential seismic source dimensions, and hence, maximum magnitude; (2) because of the spatial scale (dinensions) of lateral and vertical heterogeneity in deviatoric stress and in heterogeneous physical properties of the bedrock beneath the reservoir, there are only small potential seismic source areas (of less than 1 km square) for fault l
movement during any single seismic event; (3) the overall rate of seismicity is declining, suggesting that the stored elastic strain being relieved through the occurrence of shallow seismicity is not being
l
' replenished; (4) the evaluation has revealed that the effects of the reservoir impoundment are very limited in spatial extent (laterally as well as vertically) with a stress barrier surface beneath the seismicity; this barrier will prevent a fault from propagating through it from above or below. This limits the maximum vertical fault dimension (hence magni-tude) at shallow depths and prevents a deeper fault from reaching the surface. The occurrence of shallow seismicity under these conditions is highly unlikely to increase the probability of a larger tectonic event (intensity VII) occurring beneath the site. Alexander Testimony, follow-ing Tr. 728 at 11 to 12; See also Applicant Exhibit 1; Tr. 716-722.
9.
The Applicant drew the further conclusions from its analysis of historical data that: (1) near-field RIS events of M greater than or g
equal to 5.0 should not be considered in the Summer evaluation as this would imply induced earthquakes larger than the maximum tectonic earth-quake known to have occurred in the Piedmont tectonic region; (2) M L
equal to 4.5 is a very conservative upper bound for RIS anywhere in the Piednont tectonic province because only one RIS event (Clark Hill, 1974) has been larger than M equal 4.0 and it may not have been an induced L
event; (3) no reservoir as shallow as Monticello outside of an active seismic zone has RIS as large as M equal 4.0.
Id_. at 12-13.
L
- 10. Although the Staff found the Applicant's arguments reasonable, the Staff adopted the opinion of its consultant, LASL, that the maximum reservoir-induced earthquake is a M equal to 4.5 event as its official g
staff position. Exhibit I at 2-23 to 24; Tr.1067-69. Dr Newton of LASL indicated that the reservoir-induced seismicity is possibly asso-ciated with small scale anomalous features, but that the evidence is too
. weak to draw a definitive conclusion. There is no evidence of faults capable of earthquakes of magnitudes greater than M equal 4.0.
L Dr. Newton further noted that in cases where no faults had been identified, reservoirs like Monticello have not been associated with riacroseismicity.
Id. at 2-24.
- 11. Dr. Andrew Murphy of the NRC Staff differed with the Applicant and official Staff position on the magnitude of the maximum reservoir-induced earthquake at the Monticello reservoir. Dr. Murphy's primary concern is that the Applicant had not provided sufficient data to establish that the maximum dimension of the geological structures within the immediate vicinity of this reservoir is 1 km or less and that the maximum stress drop is less than 25 bars. He explained that the use of the 25 bar stress drop in the Applicant's calculations was justified on the basis of an abstract by a member of the USGS in which a 17 bar stress drop was reported for the August 27, 1978 earthquake that occurred in Ponticello reservoir.
In Dr. Murphy's opinion, the conservatism gained by the Applicant's use of 25 bars over the observed 17 bars may have been invalidated by apparently updated results which show that the stress drop for the earthquake was about 17 bars on one horizontal axis but about 90 bars on the other horizontal axis. Finally, Dr. Murphy indicated that if 3.2 km (length of the clusters of seismic activity) is taken as a source dimension and 100 bars as the stress drop, by Brune's model (1970) a magnitude (M ) 5.3 event is possible in the immediate vicinity of the L
reservoir. Staff Exhibit 2, Tr. 1063-65. Neither the Applicant nor Staff ascribed to Dr. Murphy's position.
8-
- 12. The Applicant claimed that part of its difference with Dr. Murphy arises from the stress drop associated with the M equal to L
2.8 event. The Applicant claimed that the USGS member's stress drop data yielded an average stress drop of approximately 25 bars (which it employs), and a peak stress drop of 90 bars (which Dr. iturphy employed)
(Tr.864). The A'pplicant also defended its utilization of the one km length fault area. Tr. 889-91, Applicant Exhibit 1.
The Applicant testified that the stress drops in question constitute an average of the stress before the earthquake rupture minus the stress af ter the earthquake rupture. Tr. 934 The Applicant testified that there is no unique measurement of stress drop in a given instance and that the results depend on different calculational methodologies Tr. 934-6, 971-72.
It further testified that there may be larger stress drops in a very small location along the fault surface and very low stress drops at other locations. Tr. 934. See also Tr. 991-93. Therefore, the Applicant testified that it was proper to calculate an average stress drop along the entire fault surface and not at one location. Tr. 934-35, 941-42. The Board agrees that a more reasonable and realistic estinate of the stress drop will be derived from average stresses before and after the event.
- 13. The stress drop value chosen by the Applicant was calculated by measuring the strong ground motion records in the site area and estimating the overall stress drop along the entire fault surface. Tr.
936. The Applicant believed that a 25 bar stress drop was appropriate based on historical observations. Tr. 937. The Applicant testified that in the studies where independent estimates of the stress values before
9 and af ter an earthquake.were taken, the stress drop was a maximum of 10 to 20 bars Tr. 943. According to the Applicant, its position had been supported by the fact that subsequent work done by the USGS member in question in the same manner as performed by the Applicant has yielded stress drop estimates essentially equivalent to the Applicant's. Tr.
942.
- 14. The balance of the Staff seismologists similarly did not accept Dr. Murphy's position. The Staff disputed Dr. fiurphy's position that the size of the area available for rupture is equal to the diameter of the clusters of seismicity (estimated by Dr. Murphy as 3.2 km). The Staff further testified that, whe,= one has no manifestation of a throughgoing fault, the diameters of clusters should not be used to evaluate the fault available for rupture. Secondly, the Staff observed that the preponder-ance of stress drop values calculated for events in the southeastern United States is less than 10 bars. Therefore, it had judged 25 bars to be a conservative estimate based on the values reported in the published literature and discussions held with persons engaged in such measurements of stress drop.
Staff exhibit 1 at 2-25 to 2-26, Tr. 1073-77, 1124-26, 11"_, 1221-22. This position was subsequently revised by the Staff. See paras. 122-126 infra.
The Staff accepted the Applicant's position that stress drop is constant over a range of magnitudes. Tr. 1129. As the proceeding progressed, the various stress drop estimates employed in arriving at estimates of maximum magnitude RIS events assumed less importance.
It became apparent that too much emphasis had been placed on the various maximum magnitude estimates in estimating RIS ground motion to the exclusion of equally, if not more, important factors such as the
. depth at which those larger postulated events would occur. See paras, 39, 85, 112-114 supra.
(b) Seismic Design.
- 15. As a result of reservoir-induced seismicity, new response spectra were suggested by the Applicant that account for the ground motion at the site for a near-field event of M equal 4.5.
This event L
results in a zero-period acceleration of 0.22g which exceeds the zero-period acceleration for the safe shutdown earthquake (SSE) of 0.15g for structures founded on rock. See Applicant Exhibit 1.
A comparison was made between the new spectra for 5% and 7% of critical damping with the original design spectrum for ?% damping.
It was found that in the high frequency region the new response spectra exceed the original spectral 4
curve.
In the judgment of the Staff, this comparison was justified, since in the original seismic analysis a very conservative 2% damping value was used. Subsequent to the original analysis, Regulatory Guide 1.61 was issued. Regulatory Guide 1.61 allows a 5% damping value for pre-stressed concrete and a 7% value for reinforced concrete structures in the seismic analysis. The Applicant demonstrated that the dominant response frequencies of all. the seismic category 1 plant structures are in the low-frequency range (lower than 9 Hz) except for the containment internal concrete structures.
Staff Exhibit la at 3-2; Applicant Exhibit 1.
- 16. The internal concrete structures have fundamental frequencies higher than 9 bz, which is in the frequency range where the original design spectrum falls below the new spectra values.
For this reason, the Applicant generated a new time history input motion using the Oroville,
11 -
California earthquake as a basis and scaled that up to a 0.22g peak value.
It was found that. the new floor response spectra based on the adjusted Oroville earthquake time history did not exceed corresponding floor response spectra for the facility design. This is due to the fact that the original artificial time history used in the facility design had conservatism built into it.
In the estimation of the Staff, this confirmed that the design of the interior concrete structures is also adequate. M.
- 17. Similar comparisons were made for equipment and systems to demonstrate that new spectra are bounded by the original design spectra except in some cases in the 20 to 30 Hz range. The Applicar.t perforned specific evaluations for those piping systens and equipment whose fundamental frequencies lie between 20 and 30 Hz and it was found that adequate design margins exist. The Staff reviewed this information and concurred. M.
- 18. At the insistance of the ACRS and Staff, the Applicant performed further analysis in order to demonstrate the ability of the plant to withstand a near-field magnitude 5 event.
Staff Exhibit Ib, Appendix D at 2.
As a result of this analysis, the Applicant testified that statistical estimates of ground accelerations at the site made by its consultant (taking into account observed strong motion data from the largest induced earthquakes at Monticello reservoir) revealed that, in order to equal or exceed the design accelerations, an event of M equal to 4.5 must be closer than 2 km, an event of M equal to 5.0 must be t
closer than 3 km, and an event of M equal to 5.5 must be closer than g
4 km. Since all reported reservoir-induced events with a ML greater than
. or equal to 5.0 have reported source depths greater than 5 km and site-specific data indicates that an event that large at Monticello would also be deeper than 5 km, the Applicant concluded that an event of the size suggested by Dr. Murphy would not adversely affect the Summer facility. Alexander testimony, fol. Tr. 728, at 13-14; Chen Testimony, fol. Tr. 733, at 2-3.
See Applicant exhibit 4; Tr. 839, 914. The Applicant testified categorically that the plant could withstand an event of M equal to 5.3 or higher. Tr. 999-1000.
L
- 19. The NRC Staff was in full agreement with the Applicant on the adequacy of the seismic design. Staff Exhibit Ib at 3-2.
Essentially, the Staff concluded that the design analysis and construction procedures applied to the facility provide a seismic capability significantly in excess of that required by either the original design seismic input or the revised seismic input, including a reservoir-induced seismic event up to a magnitude 5.5.
Tr. 1070. This insures the ability of the plant and vital equipment to remain functional and cool down the facility if such an event were to occur (Tr.1084) with ample margin of conservatism (Tr.
1087,1105). Dr. Murphy disclaimed any expertise or position on the effect that an event of M =5.3 would have on the seismic design of the g
plant. Tr. 1073. The Intervenor offered no testinony on this matter nor is it otherwise controverted in the record. This position was expanded and reinforced at the later hearing session. See, e.g., paras. 149-163 infra.
3.
Charleston Earthquake.
- 20. At the construction permit stage, the NRC Staff concluded that the weight of the seismologic and geologic information supported the
. proposition that the seismicity in the vicinity of Charleston, S.C.,
including the Modified Mercalli intensity IX-X 1886 earthquake, was related to structures beneath the Coastal Plain province in the Charleston area and should not be assumed to migrate outside of that region. The Licensing Board presiding at that stage agreed.
South Carolina Electric and Gas Co. (Summer Nuclear Station, Unit 1),
LBP-73-11, 6 AEC 213, 218 (1973).
Following the issuance of the con-struction permit, the then AEC contracted with USGS to perform an extensive geologic and seismologic investigation of the Charleston regicn. As the USGS investigation progressed, numerous working hypotheses evolved concerning the source mechanism of seismicity in that area. The summary of the current USGS position on this matter is contained in a December 30, 1980 letter from J.F. Devine, USGS, to Dr. R.E. Jackson, NRC, which is included as Appendix E to Staff Exhibit 1.
That letter concluded, in material part, that the concentration of seismicity in the Charleston earthquake epicentral area both before and after the 1886 event and the lack of post-Miocene faulting in the coastal plain or any evidence for localizing large earthquakes indicate that the likelihood of a Charleston size event in other parts of the coastal plain and Piedmont is very low. Consequently, the report continued, earthquakes similar to the 1886 events should be considered as having the potential to occur only in the vicinity of Charleston and seismic engineering parameters should 'be determined on that basis.
Id., Tr. 1070-71.
It continued with a recommendation that research on the sources of the Charleston and other east coast earth-
- quakes should continue if a more definitive resolution of the problem is to be obtained.
- 21. The Applicant also performed a reassessment of the impact of Charleston seismicity on the Sumer site in light of the new data com-piled by the USGS since the construction permit stage. The Applicant's assessment is contained in its Exhibit 1.
The Applicant position on the Charleston earthquake was summarized in the prefiled testimony of Dr. Alexander.
It concluded from the extensive work done 'oy USGS, evaluations of the most prominent hypotheses, the probabilities of future occurrences and the historical record of seismicity in the Charleston area, that there was no observational evidence to indicate that an earth-quake comparable to the 1886 avant will reoccur in any location except in the Charleston vicinity. Dr. Alexander further testified that a reoccurrence of such an event in the Charleston area will not generate ground motions that exceed the Summer design basis. Alexander testimony, fol. Tr. 728, at 16.
See Tr. 921-22.
- 22. The Staff reviewed the results of the USGS study of the Cnarleston region, the working hypotheses formulated as a result of that work, and the analyses of the Charleston region performed by the A]plicant.
Based on its consideration of this information, the Staff i
concluded that the position held at the construction permit stage is l
stn11 valid, namely, that there is no basis to assume that an earthquake equivalent to the 1886 Charleston earthquake is likely to occur anywhere but in the general vicinity of Charleston, South Carolina. The Staff took the position that the 1886 Charleston earthquake can be reasonably related to complex geologic structure unique to the region and in l
- consideration of the recurrent seismicity in the area should not, in developing the earthquake design basis for the facility, be assumed to occur at the Summer site. However, because a clear association between structure and seismicity has not been demonstrated, it recommended that geological and seismological research be continued in the Charleston a rea. Staff Exhibit 1 at 2-38 to 39.
It reaffirmed this position at the hearing that there was no basis to migrate the Charleston earthquake to other parts of the coastal plain or Piedmont provinces. Tr. 1063, 1070-71, 1155. On the weight of the authority and reanalysis supporting this view the Board agrees and reaffirms the decision reached at the construction permit stage on this natter.
4.
Wateree Creek Fault 23.
Subsequent to the impoundment of the Itonticello Reservoir and the ensuing increase in local seismic activity, the USGS contracted the services of Dr. Donald T. Secor, Jr., Department of Geology, University of South Carolina, to conduct an intensive geologic investigation of the general area surrounding the reservoir. During the course of the investigation, Dr. Secor mapped a previously unrecognized fault within the Chapin quadrangle which he named the Wateree Creek Fault. Alexander testimony, fol. Tr. 728, at 16-18. The Applicant testified that, on the basis of its review of the findings of Dr. Secor to date, the fault had been traced northward to a point approximately 2 km northeast of Peak, South Carolina and there was no observational evidence of northward continuation of the fault to the vicinity of the Monticello Reservoir.
Further, there was no geological evidence to suggest that the fault is capable nor has any seismicity been associated with it.
Id. at 18-19.
. Accordingly, the Applicant did not believe this fartor was of concern to the safety of the facility. M.
- 24. The Staff took the position in the Safety Evaluation Report that, on the basis of the information then known, it was reasonably assumed that the Wateree Creek fault may be presently adjacent to the Monticello Reservoir, that there is no historic seismicity associated with that fault, and that there is no geological evidence for capability of the Wateree Creek fault. Staff Exhibit 1 at 2-26 to 2-27.
Thus, the Staff concluded that the Wateree Creek fault did not pose a hazard to the stie. The Staff did consider it prudent, however, for the Applicant to continue to monitor the ongoing mapping of that fault. M.at2-39.
25.
In order to explore the matter further, the Board sought the appearance of Dr. Secor to explain the state of knowledge about the Wateree Creek fault. Dr. Secnr explained the status of his mapping efforts and testified that he found no evidence that the Wateree Creek fault extended into the Monticello quadrangle (Tr. 793), that the fault had not moved in " roughly one hundred million years" (Tr. 794) and that the attitude of the fault was not "particularly favorable for reactivation" (Tr. 796). Dr. Secor professed his general agreement with the conclusions drawn by the Applicant from his work. Tr. 795.
Dr. Secor further testified that there was no unknown area that would cause him to have reservations about the Wateree Creek fault upsetting the conclusions of the USGS or NRC Staff so far concerning the site.
Tr. 799.
Finally, Dr. Secor testified that the reservoir-induced seismicity was unrelated to the Wateree Creek fault (Tr. 801) and that
. the Wateree Creek fault.would not likely be activated by reservoir-induced seismicity (Tr. 803).
- 26. The Applicant felt that Dr. Secor's testimony strengthened its earlier testimony. Tr. 980.
In addition, on the strength of Dr. Secor's testinony, the NRC Staff expressed less certainty about its earlier position on the possible northward continuation of the Wateree Creek fault. Tr. 1063.
It observed that the fault was older than it had previously thought. Tr. 1092. On the basis of the entire evidence on this matter the Board finds that the Wateree Creek fault poses no hazard to the Summer site.
5.
Continued flicroseismic Monitoring.
- 27. The final seismic issue concerned the Intervenor's contention that seismic monitoring should continue until the end of 1983. The NRC Staff indicated its intention to impose a license condition whereby the Applicant must continue to mnnitor seismicity until the cnd Of 1982 and may not terminate such program unless prior written approval is received from the Staff. Staff Exhibit ib at 18-3.
The Staff believes that this continued monitoring is desirable in order to detect any possible event larger than that experienced to date. Staff Exhibit 1 at 2-23; Tr. 1069.
I The Applicant testified that it believed that the largest or approxi-mately largest reservoir-induced event had already occurred at the site--magnitude 2.8.
Tr. 886-888, 909. The Board agrees with the Staff that continued microseismic monitoring at least through 1982 is desirable and is satisfied that it will not be terminated prematurely. The Board sees no basis to impose an absolute requirement that monitoring continue through 1983.
- A.
Subsequent Developments
- 28. During the July 17 hearing session, the Board expressed certain concerns regarding the scope o' the Staff seismic review and aspects of previously heard testimony on reservoir-induced seismicity and its effect on the adequacy of the Sumer seismic design. Tr. 3790-3817. The Board concerns were in the following areas:
(1) the basis for the Staff selection of a magnitude 4.5 maximum reservoir-induced event over the recomr.endations of the ACRS; (2) selection of the peak accleration value used in derivation of reservoir-induced response spectra for seismic design purposes; and (3) the appropriateness of the Applicant's ground notion model and application of response spectra.
- 29. The Board indicated an intention to retain its own witnesses to address independently these and related concerns which it confirmed in a July 27 conference call.3_/ The Staff sought unsuccessfully to cuestion the propriety of this decision before the Atomic Safety and Licensing AppealBoard.S 30.
In an attempt to address the Board concerns without resort to independent experts, the NRC Staff filed supplemental testimony on September 15, 1981. The Board declined to admit this testinony at the September, 1981 hearing session. An updated version of this testimony was eventually admitted into evidence during the January 11, 1982 hearing session.
- 31. The Board distributed the joint report of Drs. Joyner and Fletcher and the individual report of Dr. Trifunac on September 15, 1981.
Dr. Joyner supplemented his report in a November 12, 1981 memorandum.
The separate report of Dr. Luco was distributed on September 28, 1981.
- 32. On October 20, 1981, the Board received a " Board notification" memorandum from the Staff informing the Board that it had learned from the Applicant that the USGS had recently made available unprocessed accelerometer data. This new data contained a record whose peak instrumental acceleration and response spectra at high frequencies were greater than those previously recorded near the Monticello Resevoir.5_/
The Board was also informed of the Applicant's pians to conduct field tests to evaluate the hypothesis that ground motion recorded at the accelerometer on the dam abutment at Monticello may be amplified due to topography or other factors.
C.
January 1982 Hearing Session
- 33. The second and last session devoted to seismic matters was held from January 11-16, 1982.
Several procedural issues which were first raised during a conference call between the Board and the parties on January 4, 1982 were discussed at the outset of the hearing.
Tr. 4685-93. The Board opined that expert witness-to-expert witness, rather than counsel-to-expert witness, questioning would be advantageous because of the complexity of the seismic matters-Consequently, the
-3/
These consultants were Drs. W. Joyner and J. Fletcher, USGS, Dr. M. Trifunac, University of Southern California, and Dr. J. Luco, University of California at San Diego.
-4/
See Staff certification motion, dated August 7, 1981, as supplemented August 27, 1981. Despite its agreement with the Staff that this Board had not adequately justified its retention of independent witnesses, this motion was denied by the Appeal Board in the interests of expedition. See Appeal Board Order, dated October 19, 1981, andMemorandum(ALAB-663),datedDecember14, 1981.
- Board preferred to allow Applicant, Staff and Board expert witnesses to ask questions directly of one another.
- 34. Neither the Applicant nor the Staff objected to the parties using technical interrogators under limited circumstances. Knotts, Tr. 4687-88; Goldberg, Tr. 4690. Both counsel did express concern, however, about the propriety of a procedure under which the Board's witnesses would interrogate the Staff and Applicant witnesses and thereby assune the posture of a party-advocate. Tr. 4687-4691. The Board, nonetheless, permitted the consultants to examine the Staff and Applicant panels and to offer incidental testimony during the questioning.
Tr. 5075-77, 5151.
- 35. The Board initially decided that the consultants' reports would be received as unmarked exhibits, but would be bound in the transcript.
Tr.4694,4696,4697.N Later the Board permitted the experts to adopt their reports as part of their testimony Tr. 4732-35, 4738. By the consent of the parties, Dr. Fletcher was excused from appearing personally to testify. His joint report with Dr. Joyner was introduced through Dr. Joyner.
1.
Board Consultants' Reports a.
Joyner-Fletcher Report.
(i) Maximum Magnitude
- 36. This report was limited to examining estimates of magnitude and ground motion for reservoir-induced seismicity (RIS) already in the
-5/
The memorandum was accorpanied by USGS-Open-file Report 81-1214 containing processed accelerometer data from an October 16, 1979 reservoir-induced event with peak accelerations of 0.35g, 0.369 and 0.189 for the two horizontal and one vertical components and durations of less than one second.
- I record. Joyner-Fletcher report (fol. Tr. 4696); Tr. 4749-50. With regard to magnitude, these witnesses made several points. First, the root-mean-square (rms) stress drop calculated from the August 27, 1978 earthquake should be 32 and 25 bars. Report at 2.
Dr. Joyner estimated that the October 16, 1979 event had a 60 bar rms stress drop. Tr. 4824.-
Second, scaling up the accelerometer records for the August 27, 1978 event using a 30 bar stress drop led to an M = 4.3 event; scaling up the L
October 16, 1979 event using a 60 bar stress drop led to an M = 4.5 L
event. Tr. 5388. Third, one should not put too much trust in using probabilistic frequency-magnitude techniques in predicting maximum magnitude. Report at 3.
Fourth, a more promising approach (as has been done by the Applicant and the NRC Staff) is the examination of all other cases of reservoir-induced earthquakes in the Piedmont tectonic province.
Jd. Fifth and finally, the authors note the results of work done by Mark Zoback, a colleague at the USGS, who has interpreted stress measurenents in the vicinity of the reservoir as indicating that the large stress drop events may be restricted to the upper few hundred meters. This would result in a maximum magnitude substantially less than 4.6.
Jd. Given the fact that the largest historical earthquake suspected of being reservoir-induced was approximately M = 4.3, Dr. Joyner ultimately L
6/
The reports of Drs. Joyner-Fletcher, Trifunac and Luco follow transcript pages 4696, 4704 and 4731, respectively.
- agreed with the Staff that an M = 4.5 reservoir-induced event was the g
approximate probable maxinum magnitude event. Tr. 4742-43.
- 37. Dr. Joyner's only major area of disagreement with the Applicant was that he remained " unconvinced" that the possibility of reservoir-induced events of magnitudes greater than 4 occurring at a source distance less than 1 km from the site can be excluded. He conceded, herever, that such an event was unlikely. Tr. 4699, 4742. He selected a maximum source dimension of 2 km in arriving at his magnitude estimates.
Report at 2.
Dr. Joyner questinned the sample size underlying Dr. Nuttli's hypothesis that a M 4.0 could not occur shallower than 2 r
L km in tht. eastern U.S., a concededly improbable event. Tr. 4743-44, 4808.
Dr. Joyner acknowledged that he is not an expert in the geology or seismology of the eastern United States (Tr. 4748) and that he was in a "very poor position" to argue with Dr. Nuttli on the question of eastern U.S. earthquakes in general (Tr. 4743,4809).
- 38. Without endorsing the approach utilized by Dr. Joyner to arrive at an estimate of the maximum RIS event, the Board finds that the estimates generated by Dr. Joyner are in approximate agreement with the Staff. Dr. Joyner's estimate of maximum magnitude assumes a maximum source size of 2 km. Dr. Joyner admits he is not an expert in Piedmont geology or seismology. The Board finds the substantial evidence from witnesses with such expertise suggesting that any earthquake occurring in the shallow zone underlying Monticello Reservoir will have a maximum source dimension of about 1.0 kilometer.
See, e.g. Tr. 5011-28, 5064, 5089-5114; paras. 8,12, 73, 94 supra. Utilization of this smaller l
l
23 -
source dimension would result in considerably smaller estinates of maximum magnitude than those arrived at by Dr. Joyner.
39.
In any event, as the testimony has evolved, the Board has come to the view that too much emphasis has been placed on differences between various estimates of magnitude.
It is uniformly acknowledged that much larger magnitude events are u..tkely to occur. The essential elements in terms of our evaluation of the safety significance of RIS events is not magnitude, as such, but the ground motion and associated response spectra derived therefrom. An equally, if not more, important factor in predictions of ground motion is the hypocentral distance from the site to the source of the earthquake (depth) at which such an event would occur.
As reviewed later, each witness who addressed this topic, except for Dr. Joyner, believed these postulated larger magnitude RIS events would occur at normal tectonic depth. See paras. 74-75, 113-14 supra.
(ii) Ground Motion
- 40. With regard to the subject of ground motion estimates, the report makes the following points. First, the authors have their own approach for estinating ground motion by which response spectra are anchored to peak particle velocity. Dr. Joyner testified that their approach was logically parallel to that used by the Applicant. Tr. 4698.
Second, in contrast to the Applicant, the authors' ground motion estimates include higher frequencies which nay or may not be significant to the structure. Third, the authors believe that peak acceleration saturates at a closer distance lone source radius) than that proposed by the Applicant (two source radii). Fourth and finally, the authors calculate the peak acccleration and peak velocity for nagnitudes 3.7, 4.3
. and 4.6 by scaling the observed peaks from the August 27, 1978 event.
They use simple scaling factors derived in other studies, particularly Joyner and Boore. They predict peak acceleration of 0.48g, 0.f 2g and 0.739 for these magnitudes at a distance of 700 meters. The peaks would occur at high frequencies. Report at 3-5; Tr. 4700.
- 41. Dr. Joyner explained that he used peak particle velocity to estimate spectral response values because he believed it was the appro-priate ground motion parameter for close-in ground motion (Tr. 4700) and for predicting daniage (Tr. 4761). See Joyner memorandum of November 12, 1981, fol. Tr. 4696, at 2.
Dr. Joyner testified that it was feasible to anchor response spectra to velocity because, for engineering purposes, the spectral value is important and not the manner in which it is derived. Tr. 4715-16. He acknowledged that this was a departure from the tradition of anchoring response spectra to acceleration.
I_d.
- 42. According to Dr. Joyner, velocity is probably more useful than acceleration values for estimating spectral response values in the present case because the 2 to 9 Hz frequency range. considered critical in the Staff SER, probably occupies the flat portion of the velocity response spectrum for earthquake magnitudes he regarded as relevant in the present case. Joyner memorandum, dated November 12, 1981, fol.
Tr. 4696, at 2.
He believed that response values in that flat portion should be estimated using peak velocity. M.
- 43. Dr. Joyner further testifed that his estimates of peak horizontal acceleration and velocity could be used to estimate spectral response values with the aid n' appropriate amplification factors. M.
He consid2 red the development of aopropriate amplification factors a
- proper task for the Applicant, rather than himself.
,I d.
In that regard, Dr. Joyner initially regarded an amplification of two at the site as reasonable. Tr. 4703. He later reversed this position citing uncertainty about the topography of the accelerometer site. Tr. 5395-99.
- 44. On the topic of acceleration values, Dr. Joyner testified that the different digitization rates and cut-off frequencies utilized by him and the Applicant to arrive at peak ground acceleration values from recorded ground motion records did not constitute a significant issue.
Tr. 4807-08.
In his report, Dr. Joyner had disagreed with the Applicant's use of a low digitization rate, without realizing that approximately the same results could be obtained if higher sample rates were used. Tr. 4717, 4718, 4772. While he believed high sample rates yield more accurate results and that records with high frequency content should be digitized at high rates, it is not necessary that the same digitization rate be used to determine peak acceleration and response spectra. Tr. 4717-20, 4723.
- 45. The Board is reluctant to accord much weight to Dr. Joyner's ground motion estimates for larger postulated RIS events. The Board also questions the efficacy of scaling up the observed peaks of the smaller RIS events experienced at Monticello Reservoir to date to arrive at estimates of peak acceleration for larger postulated RIS events which would occur at greater depths.
See, e.g. para. 85. Moreover, a key assumption made by Dr. Joyner in scaling recorded events to higher magnitudes is the use of a fixed depth.
In contrast to Dr. Joyner's judgmental assumption of depth, there is considerable evidence from
. Dr. Nuttli and the Staff establishing that the larger RIS events will occur at greater depths. See, e.g. paras. 8, 75, 85, 113-14 supra.
Smaller ground motion estimates would result from using normal tectonic depths for these larger postulated events.
- 46. Dr. Joyner was not aware of any damage to engineered structures cr.used by earthquakes in the range of magnitude 3 to 5.
Tr. 4787. He testified that the significance of high frequencies to a structure is an engineering judgment that he was not competent to make. Report at 5; Tr. 4772.
b.
Trifunac Report (i) Probabilistic Analysis
- 47. Dr. Trifunac did not make an independent estimate of magnitude for the maximum RIS event. Dr. Trifunac did not disagree with the Staff position that a magnitude 3 event would occur at depths less than 2 km and a magnitude 4 event at depths greater than 5 km, but he did not find it conclusive. Tr. 4841-42. Throughout his oral testimony, Dr. Trifunac frequently emphasized that, based upon his calculations, background seismicity and not RIS was the key factor to be considered at Summer.
Tr. 4705, 4735-37, 4848-46. The primary thrust of Dr. Trifunac's report was concerned with probabilistic predictions of exceeding the seismic design of the facility.
- 48. Dr. Trifunac posited that the desirable range of exceedance probabilities for " acceptable" design levels for nuclear facilities is 5 to 10 percent.
Trifunac report, fol. Tr. 4704, at 3.
According to Dr. Trifunac, assuming that (1) peak acceleration is an acceptable single parameter to employ for scaling of the design spectra and (2) the rough i
- distribution shown in hi_s Figure 1, which is based on the data in the western United States, can be employed as an approximation in the eastern United States, this figure suggests the following:
(1) the 0.15g for SSE acceleration is acceptable for M=4.5 at the site and possibly acceptable for MMI=VI and M=5.0 at the site and (2) neither 0.15g for SSE nor 0.10g for OBE peak accelerations are acceptable for 411=VII or M=5.5 at the site. M.
- 49. Dr. Trifunac stated in his report that plant design should not be,iustified by one or two stress estimates in the area because of the limited knowledge regarding source mechanisms and stress drops, (particularly in the eastern U.S.) and systematic differences between eastern and western U.S. ea'rthquakes.
Id. at 7-8.
Accordingly, he concluded that a deterministic approach, or the use of isolated measurements and case studies of several earthquakes, probably would be an unreliable or undesirable method to estimate the upper bounds of stress drop in the eastern U.S.
M.at7-9. This led Dr. Trifunac to employ a probabilistic approach to analyze distribution functions of seismicity, spectral amplitudes and earthquake occurrence for the Sumer site. He devised a uniform risk spectrum (URS) of spectral amplitudes which purportedly reflect the geometrical distribution and relative
" strength" of all earthquake sources surrbunding the site.
_Id. at 10.
- 50. Dr. Trifunac's comparison of the OBE and SSE amplitudes with URS spectra revealed that the significance of RIS decreases as the cut-off intensity increases from VII to X for the entire region. M.at
- 13. Thus, as long as the largest reservoir-induced earthquake is less
- than intensity VII or VIII, background seismicity plays a dominant role in contributing'to the URS amplitudes.
_I d_.
- 51. Dr. Trifunac acknowledged that other analysts could reach equally justifiable conclusions based on different input data and that his findings should be viewed in that light. Tr. 4898, 4951-53. He further noted that the expert who performed the probabilistic studies for the Applicar.t was no less qualified to do so than himself. Tr. 4951.
- 52. Dr. Trifunac's probabalistic analysis did not rule out the possibility of an event like the Charleston earthquake occurring near the sita. Tr. 4710-11, 4901-04. Tr. 4936, 4940, 4941-48. Dr. Trifunac intentionally lumped different tectonic provinces together to evaluate risk, even though combining the South Appalachian Region with the Georgia-South Carolina Region would lead to high risk spectra. Tr. 4870, 4895-99. lie acknowledged that he is not a specialist in eastern U.S.
seismicity or an expert in Piedmont RIS and had not made a detailed study of site geology. Tr. 4844.
(ii)ReponseSpectra
- 53. Dr. Trifunac agreed with the Staff that a design spectra which envelopes the spectra of events recorded at Monticello was a viable approach. Tr. 4711-12. He also explained that because ground motion estimates are never exact, one must account for-uncertainties by relying j
upon the engineers to use their professional judgment to add appropriate safety margins. Tr. 4852-53. Dr. Trifunac agreed that a Reg. Guide 1.60 spectra was inappropriate for RIS events. Tr. 4922-23.
In his report, Dr. Trifunac suggested that the anchor point of the RIS spectra should not be less than 0.15g, but did not recomend it as a firm value.
Id.
See Report at 3, 16.
i
- 54. Dr. Trifunac remained unconvinced by the work done by the Applicant which concluded that ground motion was amplified at the acceleroneter site. Tr. 4907-09.
- 55. Dr. Trifunac criticized the Applicant for adopting damping values in its structural dynamic calculations without justifying them on the basis of the structural system directly or on properly interpreted experiments on full scale structures of similar types.
Report at 14-15.
He regarded the Applicant's selection of damping values that coincide with the largest suggested by Regulatory Guide 1.61 as inappropriate because, in his estimation, the guide was not meant to apply to all engineering designs.
Id.
- 56. He explained that his criticism of damping values was generic.
In his opinion, the damping values generally used by engineers to analyze all types of structures are inappropriate because there is inadequate data to properly determine the values. Tr. 4919-21. The Board finds the structural damping values used by the Applicant are appropriate and realistic for present evaluative purposes and are consistent with NRC guidance and past practice.
See, e.g., paras.15,105 supra.
(iii) Seismic Design 57.
Dr. Trifunac did not have any specific safety concerns regarding the Summer facility based on the record and declined to make i
any safety recommendations. Tr. 4884-85. He did ne'. opw of any majcr i
structural damage to an engineered structure rd M w rom a magnitude 5 earthquake or less. Tr. 4848. According to this witncs5, the frequency of interest for civil engineered structures is predominately 1 to 10 Hz
. and for design of equipment and components frequencies greater than or equal to 10Hz. Tr. 4889, 4884. Dr. Trifunac noted that the exceedance of the Staff-approved design response spectra in the high frequency range would probably not effect civil engineered structures and systems but could effect high frequency equipment. Tr. 4884. He agreed that the engineer's approach to design should be to assure the best possible performance under all required operating conditions. Tr. 4885.
- 58. Dr. Trifunac concluded with some preliminary recommendations regarding a suggested RIS response spectra (based on tectonic seismicity) and computation of instrument response. Tr. 5839-40. These recommendations were devoid of substantiation in the record as we then observed. Tr. 6018. We, therefore, accord them virtually no weight in our decision.
Nor do we consider Dr. Trifunac's probability analysis of ground motion from tectonic earthquakes particularly probative of RIS concerns in this case. Among the major flaws in the study is Dr. Trifuna *s combination of the South Appalachian seismic zones with the South Carolina-Georgia seismic zone and his inclusion of aftershock data in postulating earthquake hazard at the Summer site. The former zone has higher seismicity. Tr. 5044.
It is appropriate to consider seismicity on a province by province basis. The combination of zones in Dr. Trifunac's work has not been properly justified and seriously l
I
l
- diminshes its reliability and usefulness in assigning earthquake risk at Summer.U Neither the Applicant nor Staff placed principal reliance on prob 6bility studies (Tr. 5120-24, 5938-40) nor for the reasons given are we inclined to accord Dr. Trifunac's analysis any weight.
c.
Luco Report.
- 59. The report of Dr. Luco addressed the estimates of ground motion and the associated response spectra for reservoir-induced events.
Dr. Luco indicated that he was not qualified to offer an independent estimate of the maximum magnitude earthquake. He based his estimation of ocak ground acceleration on the magnitude 4.0 to 5.5 range indicated by the record. Luco Report, fol. Tr. 4731, at 1.
(i) Ground Motion
- 60. Dr. Luco criticized the Hanks-McGuire method employed by the Applicant to estimate peak ground acceleration because the estimates of peak acceleration obtained by that method were dependent on parameters, (including magnitude, hypocentral distance, stress drop, and cut-off frequency), which may vary to a significant degree and thus render estimates of peak acceleration highly uncertain.
Id. at 1-2.
In addition, for magnitudes larger than M =4.5, the site may be located in L
the near-field in which case the model may not apply at all.
_Id. at 2.
Consequently, Dr. Luco would not rely exclusively on the Hanks-ficGuire approach, but would instead use it in conjunction with other estimates
-7/
In Public Service Co. of New Hampshire (Seabrook Station, Units 1 and 2), ALAB-667, 14 NRC
, Slip. op. at 34 (March 3, 1982) the Appeal Board found a similar probabalistic approach of questionable value in assessing site-specific risk.
. based on recorded data and available correlations.
_I d_. The Board finds that the Staff has taken such an approach.
- 61. Dr. Luco disagreed with the Applicant's use of a 25 bar stress drop to estimate peak acceleration. He argued that the stress drop para-meter in the Hanks-McGuire method has no relation with the stress drop determined by standard seismological methods and the formula used by the Applicant was incorrect.
Id. at 3-5.
According to Dr. Luco, the data from Monticello, the Hsinfengkiang reservoir in China and from California indicate that a stress drop of the order of 100 bars should be used in conjunction with the Hanks-McGuire method to obtain estimates of peak accelerations. This 100 bar stress drop would result in calculated peak accelerations of 0.36g and 0.599 for earthquakes of magnitude M =4.0 and g
4.5, respectively, at a distance of 2 kilometers.
Id. at 5-6.
Dr. Luco admitted that it was understandable that different analysts could arrive at different estimates of stress drop because of the distance and magni-tude assumed, the peak accelerations matched and the treatment of the corner frequency. Tr. 5604-06. The Board does not find it essential to reconcile all conflicting stress drop estimates in reaching an ultimate evaluation of the ability of the plant to withstand anticipated RIS events in light of other factors contributing to ground notion estimates and related response spectra.
Nonetheless, the Board finds that the current Staff 50 bar stress drop estimate represents an intermediate and reasonably conservative estimate of stress drop for estimating magnitude of RIS activity. See paras. 124-126 infra.
- 62. Dr. Luco's preliminary observation, based upon his stress drop calculations and the data from the Oroville sequences of 1975, was that
- peak acceleration levels lower than 0.259 probably could not be justified. Report at 6.
He further recommended that the peak acceleration level for Summer should be selected on the basis of the best estimate of the maximum magnitude earthquake and hypocentral distance.
Id. The Board finds that the Staff review has done this.
63.
Provided the assumptions regarding depths at which predicted magnitude events could occur are valid, Dr. Luco stated that he agreed with estimates of associated ground motion made by the NRC Staff in its supplemental testimony. Tr. 4728, 5595-96.
(ii). Response Spectra
- 64. With regard to estimates of response spectra, Dr. Luco concluded that the spectra for the August 27, 1978 M = 2.8 earthquake L
exceeded the OBE spectrum for frequencies above 10 Hz and that this exceedence would be even more pronounced for larger magnitude events.
Report at 8.
Dr. Luco was critical of the Applicant's comparison of a 2 percent damped SSE spectrum with 5 and 7 percent damped spectra for magnitude 4.5.
He regarded this comparison as inappropriate to study the effects on equipment at the lower levels of the plant.
Id. According to Dr. Luco, both spectra should have the same damping for a proper comparison of the assumed motions at foundation level. Luco report, Table 5.
Thus, if a 2 percent damping value is used in generating basement floor response spectra and compared to a 2 percent damped SSE floor response spectra, the approach would be consistent. Tr. 4970.
In his opinion, the 5 percent damped spectrum for a magnitude 5.0 event possibly could exceed the SSE 2 percent camped spectrum in the frequency range above 2 Hz and, thus, detailed review of structural elements and
. equipment with high fundamental frequencies was necessary. Luco Report at 8-9.
- 65. Dr. Luco found that the response spectra for events in the shallow region (less than 2 km), assuming a maximum magnitude of 3 at such depth, would not exceed the envelope of observed spectra. Tr. 4728.
He also testified that the Applicant's RIS spectra also seemed reasonable if a maximum magnitude event on the order of M = 4.5 were assumed at a L
hypocentral distance greater than 6 km.
Id.
If the assumptions are incorrect, Dr. Luco testified that the RIS spectrum could be exceeded by more than the 30% margin suggested by the Applicant. Tr. 4728-29.
- 66. Dr. Luco testified that the use of actual records to generate site-specific spectra was a valid means to assess the credence of ground motion estimates. Tr. 4982. He found the comparisons of Oroville and Mammoth Lake sequences in reasonable agreement with the Applicant's RIS spectrum for an M = 4.5 event. However, he observed that the average L
hypocentral distance used at Mammoth Lakes was 6 km versus the 2 km distance for the RIS spectra. Tr. 4968.
To accomodate this discrepancy, Dr. Luco would scale the spectra for an M = 4.5 event at a hypocentral L
distance of 2 km from the Mammoth Lakes data by a factor of two or three.
Tr. 4969.
- 67. Dr. Luco did not fully agree with the Applicant's position that t
conditions at the accelerometer site amplify ground motion by a factor of approximately 1.5.
Tr. 4729. He was of the opinion, however, that some reduction in ground motion for structures founded on rock versus that recorded in the free-field at the accelerometer was possible because of the scattering of ground motion waves at the foundation level, but did I
- not believe the Applicant had adequately demonstrated it. Tr. 4983-84, 5600-03, 5609. He did state, however, that standard soil-structure interaction analyses, including embedment of the foundation, would show reduction in ground surface motion in the intermediate and high frequency ranges. Tr. 5600-01.
(iii) Seismic Design
- 68. Assuming again the validity of the Staff assumptions regarding the magnitude-depth relationship, Dr. Luco was not concerned about the ability of structures to withstand estinated ground motions. He was concerned about the sensitivity of certain equipment to high frequencies.
Tr. 4730, 4973, 5595-96. He suggesced that tests or detailed calcula-tions be performed to demonstrate the safety of such equipment subject to high frequency motions.
Id. Dr. Luco was unfamiliar with the test program employed by the Applicant to qualify electrical and mechanical equipment at the plant. Tr. 4988. He acknowledged that exceedance in the high frequency range did not necessarily imply failure of equipment and components. Tr. 4988. He knew of no case in which a magnitude 5 event or less caused damage to engineered structures. Tr. 4963.
- 69. The Board finds a general convergence of opinion between Dr. Luco and the Staff regarding ground motion estimates and associated response spectra. Dr. Luco's major concern, in the final analysis, was the implication of spectral exceedances in the high frequency range upon certain equipment sensitive to such motions and recommended appropriate equipment response tests. This is a concern shared by the Staff and acknowledged by this Board. Apart from the considerable evidence about the design adequacy and conservatism of such equipment, a detailed
- analysis to confirm the requisite design margins for postulated RIS events is already in progress.
See, e.g., para, 162. This is directly responsive to Dr. Luco's residual concern with regard to essential safety equipment.
2.
Applicant's Affirmative Case
- 70. The Applicant's additional seismic testimony was presented by several panels of witnesses. These panels included previous witnesses Dr. McGuire, Alexander, McWhorter, Talwani and Chen and new witnesses Drs. John A. Blume, Malcolm R. Sommerville (URS/ John Blume and Assoc.),
Otto W. Nuttli (St. Louis University) and Geoffrey R. Martin (ERTEC).
- 71. These witnesses addressed the issues raised by the Board's con-sultants and the newly-discovered accelerometer data and presented new information and analyses in support of positions taken by the Applicant in earlier hearing sessions. The subjects discussed by the Applicant's witnesses can be divided into three principal areas: maximum magnitude; ground motion; and seismic design.
In addition, the Applicant's wit-nesses evaluated the reports presented by the Board's consultants.
a.
Maximum 11agnitude
- 72. The Applicant submits that M( = 4.0 is the appropriate maximum magnitude event that can be induced by the Monticello reservoir.
Alexander testimony, fol. Tr. 5028, at 12; Tr. 5011. Several lines of evidence were presented in support of this conclusion.
73.
First, historic experience both locally and within the Piedmont province was cited. Dr. Alexander testified that no reservoir not associated with active fault zones has produced significantly large earthquakes. Tr. 5011.
In the Monticello region of the Piedmont
37 -
province, all events thought to be induced by reservoirs have been less than 11 = 4.0 with a single exception. That exception, the Clark Hill L
event with f1 = 4.3, was said to be "of questionable association with the L
reservoir itself" because it occurred long after reservoir impoundment and because other comparable events had occurred in that region prior to impoundment. Tr. 5011-12. The data base for this conclusion that RIS events in the Piedmont province are less than M = 4.0 consisted of 59 L
reservoirs of similar size to Monticello representing about 2200 years of reservoir operation. McWhorter testimony, fol. Tr. 5031, at 1-2; Tr. 5029.
74.
Second, several types of site-specific evidence were said to support the M = 4.0 for RIS conclusion. The spatial extent of RIS at L
Monticello is confined laterally to the immediate area of the reservoir and vertically to a depth of less than three kilometers.
In fact, with respect to the vertical aspect, Dr. Alexander testified that over 98% of the events were shallower than two kilometers and approximately 80%
shallower than one kilometer. Tr. 5012. These bounds were reached quickly (within one and a half years) and have not expanded since that time. Tr. 5012-13. Only microearthquakes (less than M = 3.0) have L
occurred since impoundment and their average rate of occurrence has been steadily declining with time. Tr. 5013.
The shape of the observed frequency versus magnitude curve is consistent with a limiting magnitude.
Tr. 5013. And, finally, heterogeneities in rock properties which have been documented were said to limit the extent of any single fault move-ment to a distance estimated to be one kilometer. Tr. 5014. Recorded
. stress levels at Monticello were also said to be not sufficiently high to allow extensive ruptures to occur very deep. Tr. 5019.
- 75. The Applicant's consultants also examined worldwide reservoir-induced seismicity to evaluate the conclusion that the maximum RIS event at Monticello would be M = 4.0.
That examination showed that out of 64 L
confirmed cases of worldwide reservoir-induced seismicity, only 11 had induced events of magnitude 5 or greater. Of those 11 cases, nine were associated with active faulting and the other two were most likely associated with active faulting. McWhorter, Tr. 5029. With the possible exception of the New Madrid, Missouri earthquakes in 1811 and 1812, there have been no observations of surface rupture occurring in the eastern United States. Nuttli testimony, fol. Tr. 5164, at 5.
No active faulting exists at the Summer site. See paras. 23-26, supra. The Applicant's experts drew a conclusion from this comparison that "for reservoirs in intraplate tectonic settings away from active tectonic elements, a maximum magnitude of about 4 appears to be appropriate."
McWhorter, Tr. 5030.
- 76. The Applicant further argues that the shallow depth of the reservoir-induced seismicity at Monticello Reservoir is an important factor limiting the maximum magnitude of such events. Approximately 98%
of the events have been less than 2 kilometers deep.
Experience from earthquakes throughout the entire central and eastern United States suggests that magnitude 4 is the upper limit for such shallow earthquakes.
Nuttli testimony, fol. Tr. 5164, at p. 4; Tr. 5173, 5175.
- 77. The Applicant also conducted probability studies of seismicity and ground motion at the Summer site in order to assess the probabilities
. of occurrence of various magnitude earthquakes. McGuire, Tr. 5033.
Applicant's consultant Dr. McGuire calculated that the average annual probability that an earthquake of magnitude greater than 4 will occur at an epicentral distance less than 1 kilometer is about 1 x 10-4
- McGuire,
~
Tr. 5034-35. This implies a return period of about 7400 years. McGuire testimony, fol. Tr. 5042, at RM-4, p. 4.
An alternative calculation using data from reservoir-induced seismicity in the Piedmont region of the eastern United States provided a probability of occurrence of 3 x 10-5 or a return period of 34,000 years.
.I d.
In either case, the Applicant characterized the probability of occurrence of a magnitude greater than 4 in the vicinity of the facility as " remote." McGuire, Tr. 5035. This conclusion was supported further by probabilistic analyses for tectonic earthquakes which indicated a return period between 1700 and 10,000 years for an intensity VII tectonic event. Tr. 5036; McGuire testimony, fol. Tr. 5042, at RM-4, p. 7.
73.
To summarize, then, the Applicant's position is that M = 4.0 L
is t' e maximum magnitude event that can be induced by the Monticello reservoir. This position was supported by evidence concerning historic experience near Monticello reservoir, within the Piedmont province, and worldwide.
It was also supported by site-specific geophysical data, by the shallow depth of reservoir-induced seismicity at Monticello and by probability studies of seismicity at the Summer site. While placing limited weight on the probability studies in the present context, we find the balance of the evidence adduced quite compelling. The Board finds Dr. Nuttli's magnitude-depth relationship particularly significant.
l
b.
Ground Motion
- 79. Questions have been raised by the Board's consultants about the appropriateness of the ground motion values utilized by the Applicant in its design of the Summer facility, particularly in light of accelerometer data recorded near the Monticello reservoir. The Applicant's witnesses addressed the subject of ground motion extensively during their supple-mental presentation.
- 80. The Applicant reiterated its position that the SSE and OBE values established at the construction permit stage were appropriate and conservative. Nuttli, li. 5482.
In the first instance, this conclusion was supported by the Applicant's ground notion model. The Applicant adopted a theoretical model -.the Brune model -- which has been shown appropriately to estimate ground motions in other areas, and selected parameter values as inputs to the rodel using the observations available at Monticcllo reservoir. McGuire testimony, following Tr. 5042, at RM-3
- p. 4.
For the important parameter of seismic stress drop, the Hanks and McGuire method of estimating stress drop was used. M.atRM-3p.6.
- 81. The Applicant went beyond its reliance on the mathematical model, however, and presented other lines of evidence to support its ground motion figures.
For example, the Applicant cited recent data from Joyner and Boore on California earthquakes to support the adequacy of the SSE acceleration. Tr. 5005; Blume testimony, following Tr. 5234, at B4.
The Applicant also presented testimony from Dr. Nuttli that, for normal depth earthquakes in the eastern United States, peak horizontal accelerations in the near-source region of M =5 earthquakes are estimated L
41 -
at 0.119, which is less than the SSE value. Nuttli testimony, fol.
Tr. 5164, at 7.
- 82. The high instrumental acceleration figures (0.359) recorded by the Jenkinsville accelerometer were discounted by the Applicant on several grounds. The Applicant noted, first, that these figures are incompatible with published intensity-acceleration relations, which predict much lower acceleration values. Blume testimony, fol. Tr. 5234, at 82-3, B3-3, B3-4.
- 83. The Applicant also argued that the Jenkinsville accelerometer reading did not reflect the true free-field values because of the dynamic effects of instrument site topography or soil cond,itions. Blume testimony, fol. Tr. 5234, at 37, B5-1.
This argument -- referred to as amplification -- was supported by the Applicant's active field experiments. Through two explosion tests, the Applicant acquired data for a comparative study of ground motion at different sites.
Dr. Somerville presented these data and found that, in the frequency band from 5 to 50 Hz, amplitudes in the free-field on saprolite were twice those recorded in massive structure foundations on bedrock. Tr. 5497.
He concluded that this difference was substantial (Tr. 5498) and that the i
Jenkinsville acceleroneter recordings should be modified in assessing the effects of RIS on the Sumner facility, which is a massive embedded structure founded on bedrock. Somerville testimony, fol. Tr. 5169, at 1.
He attributed these observed differences to several effects which could not be resolved uniquely from the data.
Id. a,t 6.
l
- 84. Dr. Martin also addressed the subject of amplification of the Jenkinsville accelerometer readings. He testified that topographic
42 -
effects were unlikely to generate significant amplification of these records (Tr. 5525); although such an effect could occur, it was likely to be small (Tr. 5540). Dr. Martin concluded, on the basis of a test conducted on the concrete pad upon which the accelercmeter is situated, that soil-pad interaction could result in peak acceleration readings that are SN greater than the input for true free-field acceleration).
Tr. 5554.
- 85. The amplification argument was also supported by comparison to data from ground motions at Oroville dam in California (Blume testimony, following Tr. 5234, at Appendix B-5) and by data from the Hsinfengkiang reservoir in China from which a magnitude-acceleration equation was developed.(M.atAppendixB-11).
86.
Finally, the Applicant argued that it is not appropriate to scale up the ground accelerations recorded for very shallow, very small earthquakes to extrapolate acceleration values for earthquakes of ML greater than 3 because such larger earthquakes will occur at greater depths than those which have been observed. Nuttli testimony, following Tr. 5164, at 6, 8.
The Board agrees and finds this position consistent with the weight of the evidence on this subject.
c.
Seismic Design
- 87. The Applicant argued that the response spectra developed by it to represent RIS ground motion are consistent with published work, are based on appropriate data and meets the requirements of Regulatory Guide 1.60.
The shapes of the spectra utilized were taken according to the widely accepted Johnson and Traubenik figures. McGuire testimony, fol.
Tr. 5042, RM-1 at 2.
These shapes compared favorably with the standard
- Regulatory Guide 1.60 spectrum.
_I d. Further, a comparison of the spectra developed by the Applicant with those derived from other studies was said to support what was done in this case. Specifically, response spectra from Mammoth Lakes (Somerville, Tr. 5347-51) and from Oroville aftershocks (McGuire, Tr. 5352-54) were cited in support of the mean plus one standard deviation response spectrum utilized by the Applicant for the Summer facility.
- 88. The Applicant acknowledged that certain perm acceleration readings exceeded the design response spectra. However, the Applicant's witnesses presented several lines of evidence to demonstrate that these exceedances would not adversely affect the facility. These lines of evidence were divided into two groups -- generic and site-specific.
89.
In the generic grouping, the Applicant argued that high peak accelerations at high frequencies (where the exceedances had occurred) were without potential for damage. The Applicant, therefore, submitted that peak ground acceleration should not be used directly in design.
Blumetestimony,fol.Tr.5234,at14-15.8f Evidence from Project Rulison was cited to demonstrate the low energy and low damage potential of high frequency accelerations. Blume testimony, fol. Tr. 5234, at Appendix B-14; Tr. 5486-89.
In fact, the Applicant's witnesses testified that they knew of no case where earthquakes of magnitude 5 or less caused 8f The Appeal Board has consistently accepted the use of effective (mean) ground acceleration for facility design purposes. Seabrook, supra n.6, Slip op at 43; Pacific Gas and Electric Co. (Diablo Canyon Nuclear Station, Units 1 and 2), ALAB-644, 13 NRC 903, 940 (1981).
44 -
damage to engineered facilities or equipment whethe or not any seismic design was involved. Blume, Tr. 5262.
Earthquake damage is caused by low frequency, long duration shaking rather than high frequency, short duration ground motion. By contrast, specific accelerometer data from the Monticello dam abutment show that RIS earthquakes there have durations of a few tenths of seconds at most. Nuttli, Tr. 5174; Blume, 5252. Observations that the Fairfield hydro plant near the reservoir had not been damaged by these RIS earthquakes were cited in support of the Applicant's argument. Sommerville, Tr. 5178; Blcme, Tr. 5316, 5322; Chen, Tr. 5346).
- 90. The Applicant's witnesses listed several types of conservatisms in nuclear plant design generally which combined to provide safety margins far above those generally assumed. These conservatisms included the following: design values are established conservatively; material strength is specified conservatively; the assumption that both horizontal components of ground motion are equal is conservative; structural and mechanical systems are assumed to have constant natural periods; floor response spectra are used for upper levcis of the facility; response spectra are enveloped; no credit is given for ductility; and no credit is given for design for other than seismic stresses. Blume, Tr. 5247-49; Blume testimony, fol. Tr. 5230, at 28-33 and Appendix B-10.
In sum, according to the Applicant, a plant's likely capacity is greater than the specified acceleration by as much as several hundred percent."
Id. at 33 and Appendix B-15; Blume, Tr. 5511. The Applicant's witnesses also stated that damping of several types is present beyond the 7%
critical damping for which credit is given.
Id. at 35-37.
- 91. Dr. Blume cited three examples to demonstrate the ability of engineered structures to withstand earthquake damage.
In the San Francisco earthquake of 1906 (Ms=8.25), all but seven of 52 major b fildings were returned to service even though they had not been designed for earthquakes at all. Tr. 5259. An Esso refinery very close to the Managua, Nicaragua 1972 earthquake (M=6.25) was shut down for inspection for 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> and then returned to service.
Id..
Finally, the Huachipato Steel Plant in Chile was subjected to a nearby magnitude 7.5 earthquake in 1960 but was shutdown for only 6 days for minor repairs. Tr. 5259-60.
92.
In the area of site-specific conservatisms, the Applicant's witness, Dr. Chen, discussed the seismic safety of structures and equipment at the Summer station. He cited several conservatisms which contributed to seismic safety at Summer.
For example, the large foundation (or tau) effect was explained as a phenomenon peculiar to large structures analogous to the manner in which large ships " iron out" waves. Chen testimony, fol. Tr. 5324, at p. 8-9; Blume testimony, following Tr. 5234, at p. B11-2. A substantial ductility effect was identified as applicable at Summer.
Id. at 9.
A higher in_ situ strength of concrete and steel was also cited.
,Id..
Other structural conserva-tisms included larger than specified prefabricated steel member size, radiational damping, spectra enveloping, heavy equipment mass ratio effect, equipment qualification by test, and strain hardening.
Id_. at 9-11.
Dr. Chen cited the response of the El Centro, California steam plant to the 1979 Imperial Valley earthquake as demonstrative evidence that a well-engineered structure can resist seismic loads "many times
. higher than its design value."
_Id. at 11-12; see Blume testimony, fol.
Tr. 5234, at B12-2, B12-3.
- 93. Applicant's witnesses concluded that the V. C. Summer facility will not be adversely affected by earthquake ground motion. McGuire, Tr. 5001, 5003-09; Chen, 5346. Reservoir-induced seismicity at Monticello will be characterized by small magnitude, short duration ground motions which pose no seismic hazard to the V.C. Summer nuclear station. McGuire, Tr. 5000. Any tectonic earthquakes of greater magnitude (M = 4.0 to 5.0) would be at normal depths and would not cause L
damage to the Summer facility. McGuire, Tr. 5006; Blume, 5332, 5486.
The Board finds the conclusions generally substantiated by the weight of the evidence adduced at the hearing.
d.
Applicant's Evaluation of Board Consultant Reports (i) Evaluation of Joyner-Fletcher Report.
- 94. The Applicant made several points in response to the Joyner-Fletcher report. Two concerns of Joyner and Fletcher, the effect of corner frequency on the stress drop estimate for the August 27, 1978 earthquake, and the digitization of the record from that event at 500 points per second, are not issues at all. The Applicant has analyzed both in detail, and its recommendations incorporate those analyses.
Applicant Evaluation--Joyner-Flecther Report (McGuire), fol. Tr. 5042.
Dr. Joyner discounted this as a significant issue in his direct testimony. See para 44 supra.
- 95. The Applicant disputed the validity of Dr. Joyner's maximum magnitude estimates for RIS events because of his critical misassumption of a 2.0 km fault source dimension in scaling magnitude from the
' j August 27, 1978 event at Monticello Reservoir. The Applicant presented considerable evidence that the shallow heterogenecus environment beneath the reservoir implied a maximun fault dimension of 1.0 km. Tr. 5011-28, 5064-65, 5089-5114. The Applicant regarded fault length as a key determinent in magnitude estimates. Alexander testimony, fol. Tr. 5028, at 5; Tr. 5015-17.
- 96. The Applicant acknowledged that Joyner and Fletcher's concern regarding ground motion saturation involves significant interpretation and judgment. Drs. Joyner and Fletcher offer no alternative methods to determine the distance within which ground motion amplitudes are saturated, except to use the distance between the source and recording site for the August 27, 1978 event, a chance occurrence. As noted above, the Applicant offered significant expert opinion (particularly from Dr. Nutt11) regarding minimum depth of larger postulated RIS events in the Piednont.
See paras. 8, 73-74 supra. Dr. Joyner did not refute this position but simply questioned its conclusiveness.
- 97. The Applicant further criticized Drs. Joyner and Fletcher's use of a single component peak acceleration from that event's record to scale peak acceleration and make recommendations. According to the Applicant, such a procedure is without precedent.
It takes no account of important parameters such as earthquake stress drop, distance to larger events, instrument and record processing procedures, and scaling of response spectra from the predicted peak accelerations. Drs. Joyner and Fletcher state that the methods of Newmark and Hall (1969) can be used to compute response spectra given its estimates of peak acceleration (and velocity),
but the broad-band amplification factors of Newmark and Hall (1969) would
- be wholly inappropriate.for what Joyner and Fletcher admit would be high frequency notions. According to the Applicant, this illustrates a position which it has taken since the beginning:
the estimates of peak acceleration must be made in light of the overall design problem and local conditions at the facility. Jd.at12.
(ii) Evaluation of Trifunac Report.
- 98. The Applicant introduced a written evaluation critical of the Trifunac report. Applicant Evaluation--Trifunac Report (Somerville),
fol. Tr. 5169. The Applicant's evaluation makes the following principal points. According to the Applicant, Dr. Trifunac's coments imply that design or effective acceleration (the acceleration at zero period and a spectral response diagram, i.e., the acceleration used to " anchor" the design spectrum) can be compared directly with the peak instrumental acceleration. Because strong motion accelerographs may record high frequency acceleration pulses that have no effect on structures, particularly for ground motions close to the cause of the fault, the two are not equivalent. The Applicant cites approvingly to the Diablo Caiyon seismic hearings which endorsed this concept. Jd. at 1; See para. 8ts supra and n.7.
- 99. The Applicant's evaluation continued by claiming that Dr. Trifunac's commentary rests very largely on regression analyses performed by himself either individually or with associates. Wcrk by other authors on regression analyses of strong motion data is ignored.
The Applicant cited studies which refuted Dr. Trifunac's claim that the strong motion data now available are not adequate to find the form of the distribution function of peak ground acceleration.
. 100. According to the Applicant the results of regression analyses performed by Dr. Trifunac using magnitude data are similarly marred by erroneous statistical treatment.
It argued that such work should not be used in appraising peak ground accelerations for the Summer station.
101. The evaluation further noted that a principal difficulty in applying Figure 1 of the Trifunac report in assessing design accelerations is that the difference between peak acceleration and design acceleration is overlooked. According to the Applicant, if adjusted for statistical error and for the difference between design and free field instrument acceleration, Dr. Trifunac's Figure 1 would indicate that the SSE design acceleration is appropriate for ground shaking of MM intensity VII or ground motion due to an earthquake of magnitude 5 to 5.5 occurring in the immediate vicinity of the site.
102. To summarize, the Applicant indicated that the methodology usea by Dr. Trifunac in estimating peak accelerations for given intensities and magnitudes leads to overestimation of acceleration.
Therefore, it regarded his conclusions concerning the adequacy of the SSE as inappropriate.
_Id. at 3-4.
103. The Applicant also critized Dr. Trifunac's coments on stress drop analyses.
It contended that the comparisons of peak acceleration to stress drop ratio by Dr. Trifunac are invalid. The stress drops used by the Applicant are derived from rms. Those cited in Dr. Trifunac's references are determined by spectral methods which are often one-tenth the value determined by rms for the same earthquake. Thus, according to the Applicant, the discrepancy found by Dr. Trifunac is easily explained by the factor of 10 difference in stress drop estimates by various
' methods.
It does not imply that the Applicant's peak acceleration estimates are low.
104. With regard to Dr. Trifunac's probability studies, the evaluation provided at least two reasons why Dr. Trifunac found larger probabilities than those of the Applicant.
First, Dr. Trifunac used the recurrence curve of Chinnery (1979) for the southeast United States.
Act.ording to the Applicant, Chinnery's investigation had the purpose of comparing general seismicity characteristics in different parts of the eastern United States not calculating seismic hazards at individual sites. The second difference cointed out is in the attenuation curves that are used to estimate ground motion characteristics.
105. The Applicant used (for Modified Mercalli intensity) an equation based on MM intensity observed during the 1886 Charleston earthquake which is the most extensive data base available for the southeastern United States.
For acceleration, an equat ion developed by Nuttli for the central United States was used. The Applicant argued that these attenuation functions are the most site-specific, least interpretive attenuation equations available. Those purportedly used by Trifunac estimate spectral velocity as a function of earthquake intensity and distance. While this is a novel approach, there is no eastern U.S.
data with which to judge its appropriateness nor has this methodology received substantial peer review. Thus, according to the Applicant, the use of this equation to make probability calculations and statements results in highly tenuous conclusions and should be reviewed with caution.
Id. at 6.
. 106. With regard to other Trifunac comments on structural damping, the Applicant contended that the primary reason for using 7 percent instead of 2 percent damping is due to the fact that 7 percent is more realistic than 2 percent during a 0.229 near-field earthquake for structures originally designed for a 0.15g far-field earthquake and not solely because it is permitted by Reg. Guide 1.61. According to the Applicant, the 7 percent damping was verified by test data discussed extensively during the Diablo Canyon seismic hearing. The Appeal Board decision in that natter acknowledged that 7 percent damping is appropriate.
Id. at 6-7.
(iii) Evaluation of Luco report.
107. The Applicant's evaluation of the Luco report made the following points. The equations derived by Luco differ little from those used by the Applicant, and lead to an estimate of a 26 bar stress drop for the August 27, 1978 earthquake when the correct source-to-site distance is used (0.67 km). Luco's result of a 100 bar stress drop is obtained because he uses incorrect distance. Applicant Evaluation--Luco Report (McGuire), fol. Tr. 5042, at 10.
108. Similarly, Dr. Luco's characterization of earthquakes at Hsinfenkiang Reservoir with a 100 bar stress drop is incorrect. He has erroneously used surface-wave magnitude for local magnitude, and uses an upper frequency of 20 Hz whereas the instruments are linear up to 35 Hz.
Both errors result in an erroneously large estimated stress drop for the recorded events. The results provided by the Applicant in Appendix XI to the " Supplemental Seismological Investigation" (stress drops less than 25 bars) are correct. Thus, Dr. Luco has presented no new analyses or data
- to indicate that a stress drop greater than 25 bars should be used at Monticello Reservoir.
Id.
109. The predictions of peak acceleration made by Dr. Luco are invalid because they assume a stress drop of 100 bars, which is unsupported by any analysis. The peak acceleration data from Oroville aftershocks presented by Luco are biased toward the largest peak accelerations by a facter of more than two and, therefore, cannot be used to choose peak accelerations for seismic evaluations.
Further, most of the data presented by Luco are from California earthquakes where stress drops of 100 bars are common. These data are inappropriate to characterize the very shallow, low stress drop events at Monticello.
Id.
110. The response spectra developed by the Applicant cor. form to the requirements of Regulatory Guide 1.60.
They are consistent with mean-plus-one standard deviation spectra and were developed to reflect the near-field, rock site conditions of reservoir-induced earthquakes affecting the Summer station.
Id.
3.
Staff Affirmative Case 111. The Staff direct case at this session consisted of its updated supplemental testimony on seismicity (following Tr. 5758). The testimony was introduced through a panel consisting of the following: Mr. James P.
Knight and Drs. Phyllis Sobel, Robert E. Jackson (Geosciences Branch Chief), Leon Reiter (Seismology Section Leader), Pao-Tsin Kuo (Engineering Section Leader), Carl A. Newt, n and Andrew Murphy.
Dr. Sobel provided a summary of the seismological portion of the testimony. Tr. 5760-71. Mr. Knight provided a summary of the engineering portion of the testimony. Tr. 5771-75. Dr. Newton
- summarized his additional comments contained in the testimony.
Tr. 5775-78.
112. The testimony contained a definition of terms in the fields of seismology and seismic design. Supplemental testimny at 6-11.
The testimony then reviewed the procedural history of the case and the genesis of the staff review.
Id. at 11 '.2.
In addition to addressing the principal seismological issues on a topical basis, the testimony also factored in the relationship between time history and response spectra for engineering design and the numerous engineering conservatisms associated with the Summer seismic design.
Id. at 50-57, Knight, Tr. 5772. These conservatisms include: conservatisms associated with the characterization of the design event, conservatisms associated with the methodologies for seismic analysis, design, and qualification, and conservatisms resulting from actual (versus design) structural and mechanical resistance.
_Id. at 52-57.
a.
liaximum Magnitude Earthquake (1) Normal Depth 113. The Staff position in the SER was that the maximum earthquake for design purposes was M = 4.5.
The distance or depth was not speci-L fied. The Staff approved a spectra developed by the Applicant for this earthquake with the recognition that short duration, high-frequency accelerations from small events could be higher.
In developing ground motion estimates the Applicant used a model which assumed that a M = 4.5 L
earthquake would occur at a distance of 2 kilometers.
Id. at 25. The Staff regards depth as one of the key factors in estimating RIS ground motion. Sobel, Tr. 5763.
- 114. According to.the Staff, a definition of maximum magnitude to be used for design purposes is particularly difficult with respect to RIS. The Staff continues to place great emphasis on experience at other reservoirs in the Piedmont and the largest earthquake in the Piedmont that has tentatively been associated with RIS, the magnitude 4.3 event in 1974 near the Clark Hill reservoir.
It was also observed that world-wide RIS earthquakes greater than about a 4.5 occurred in active tectonic areas dissimilar to the Monticello region. Based upon this experience, the Staff adhered to the position that a maximum magnitude of 4.5 was conservative.
Id. at 25-26, 41.
115.
The Staff cited approvingly from the testimony of Dr. Nuttli who found no evidence anywhere in the eastern or central U.S. of magnitude 4.5 events occurring at shallow depths (2 kilometers or less).
While no depth has been estimated for the Clark Hill earthquake, the intensity and felt area are similar to other earthquakes of this magnitude in the eastern and central U.S. for which Hermann (1979) estimates typical depths of 5 to 16 kilometers.
Id.; Reiter, Tr. 5886.
Based on this, the Staff took the position that, if indeed a 4.5 event were to be triggered by the reservoir at Monticello, the best estimate as to its depth would be this tynical normal depth range of 5 to 16 kilometers.
Id.; Sobel, Tr. 568-69. The Board finds this position both reasonable and convincing and supported by the weight of the evidence on l
this subject.
(ii) Shallow Depth 116. The Staff next addressed its estimation of the maximum magnitude event for the shallew zone of reservoir induced seismicity
-r---
y y
wyr g-y v-y wy w
y-w
- (upper 2 kilometers). The Staff reviewed several arguments presented which would limit the maximum magnitude.
First, the maximum magnitude shallow earthquake at the reservoir to date has been about M = 3.0.
L Sobel, Tr. 5769. Second, in-situ stress measurements of M. Zoback of the USGS at Monticello tend to indicate that the events with larger stress drop should occur in the upper few hundred meters.
Id. at 27. This is supported by the fact that the largest stress drops and the highest peak accelerations have come from events which occur in the upper few hundred meters and also from the fact that seismicity decreases with depth under the reservoir. Sobel, Tr. 5769. This position was also influenced by the tendency of the frequency-magnitude curve to indicate saturation at about M = 3.0.
Id. ; Reiter, Tr. 5947, 5953.
L 117. The Staff recognized Zoback's measurements in arriving at its conclusions with regard to maximum magnitude. M. at 41; Reiter, Tr. 5891; Sobel, Tr. 5897; Newton, Tr. 5901-02. According to the Staff, Zoback's findings appeared to be borne out by the Applicant's calculation of stress drop for the strongest ground motion recorded at tionticello.
The Staff noted the Applicant's estimates of depths for these events ranging from 70 to 360 meters with the highest stress drops being associated with the earthquake of M[ = 2.4 and 2.8 at 200 and 70 meters, respectively. M.at27.
l 118. Third, there has been an overall decline in the rate of l
seismic activity which suggests that stored strain is not being replenished.
Id. at 28.
Fourth, it referred to Dr. fluttli's testimony which indicated that no earthquake greater than M = 4.0 has occured at L
such shallow depths anywhere within the eastern U.S.
Fifth, the Staff's
- ground notion estimate is not dependent on the Applicant's ground motion model which includes uncertainties in source characteristics such as stress drop and interpretations of saturation of ground motion with distance. Sobel, Tr. 5769. Sixth, the Staff has chosen to envelope recorded ground motions instead of choosing the 50th or an 84th percentile'1evel of a suite of spectra. M. The Staff position does not incorporate the 50 percent reduction factor proposed by the Applicant.
E-119. Based on the above, the Staff took the position that the best estimate of naximum magnitude for the shallow zone of seismicity is M L
3; Reiter, Tr. 5883, 5914-15. The Board finds this position both reasonable and persuasive and generally consistent with other competent testimony on the subject. The Staff noted that this is the approximate maximum magnitude that has already occurred at the reservoir. M.
120. The Staff understood, however, that rigorous demonstration that earthquakes with magnitudes between 3.0 and 4.5 will not occur within the shallow zone is beyond the state of the art. The Staff thus
' examined the sensitivity of the ground motion expected from these events given the assumption that a larger earthquake than that which has already occurred will occur. M. at 28; Sobel, Tr. 5769; Reiter, Tr. 5883.
b.
Ground Motion Estimates:
M, = 4.5 RIS (i) Hanks-McGuire Model 121. The Staff regarded the Applicant's model for estimating RIS ground motion as a reasonable approach for arriving at estimates of ground motion.
It utilizes generally-accepted source theory to incor-porate those parameters which determine ground motion, including earth-
- quake magnitude (or moment), stress drop, and the properties of the earth through which the seismic waves travel.
It is the Staff view that the model is reasonable from a theoretical point of view and has been shown to be consistent, within some band of uncertainty, with a wide range of strong motion data in California. Ld.at29.
122. The Staff noted that the principal criticism concerning the Applicant's model has centered around the choice of input parameters, particularly stress drop, to be used in estimating ground motion and the choice of the highest cut-off frequency (fu) passed by the strong motion instrument or the earth itself.
Id. at 31. This latter parameter was addressed by Drs. Joyner, Fletcher and Luco in their reports and by the Applicant in response to these documents.
123. According to the Staff, with the availability of high quality instrumentation and filtering techniques, the value of fu is determined by the choice of the filtering and digitization procedure.
In addition, deriving the rms stress drop requires the evaluation of an integral.
Different approximations have been suggested which deal with the very high frequencies differently. The choice of the highest frequency is only of real significance when dealing with small earthquakes and relatively low values of fu. Id. At Monticello, a large part of the energy from small nearby earthquakes occurs at very high frequency.
,I d_.
at 32; Sobel, Tr. 5912.
124. The estimates of stress drop arrived at by the Applicant and Joyner and Fletcher for the August 27, 1978 magnitude 2.8 event are relatively similar when tne higher frequencies are included but may be significantly different if they are not. Utilizing this earthquake,
- which yielded the strongest ground motion known at Monticello prior to discovery of the October 16, 1979 recording, the Applicant assumed that the appropriate stress drop to be used in estimating ground motion for larger earthquakes was 25 bars. M.at32.
125.
In the SER the Staff found the Applicant's choice of 25 bars is a reasonable rms stress drop to be used in estimating RIS ground motion. Based on new information which has come to light in the past few months, the Staff reexamined its position. The most significant aew information are the additional records of ground motion at Monticello made available by the USGS.
126. The Applicant has calculated the average rms stress drops for all six of the events recorded thus far at Monticello (magnitude 2.2 to 2.8) which have significant ground motion. These rms stress drops are approximately 12, 19, 23, 42, 7 and 48 bars.
Forty-eight bars is associated with the October 16, 1979 M = 2.8 earthquake which also had L
the highest peak acceleration.
Id. at 32-33.
127. The Staff noted the Applicant's arguments that using moment magnitude will reduce the maximum rms stress drop to 37 bars. The Staff testified that, since its determination of maximum magnitude was on the basis of the local magnitude (M ) scale, it thought it more appropriate L
to use that scale. M.at33.
128. The Staff noted the Applicant's additional argument that, because of soil amplification at the site which recorded the strong motion records (dam abutment), the estimate of stress drop should be reduced by 40 percent. The Staff did not ascribe to this position. M.
at 33. The Staff also cited work done by other seismologists, including
. Drs. Nuttli and Fletcher, regarding stress drop measurements. The Staff noted the difficulty in detennining with certainty what rms stress drops should be used in estimating ground motion for an M = 4.5 event near L
Monticello since there is no stress information at normal tectonic depths. However, based on its reexamination of the additional information, and to account for observed variations and possible increase with magnitude, the Staff concluded that 50 bars (the approximate maximum observed thus far at Monticello) is the appropriate rms stress drop to be used in estinating ground motion for an M = 4.5 earthquake.
Id. at L
33-34.
129. Overall, the Staff concluded that the Applicant's model is physically reasonable but needed to be treated with caution and, when possible, used in conjunction with other approaches. The Board agrees.
The Staff testified that for small magnitudes and for distances within several kilometers, the amplification factors utilized by the Applicant in deriving response spectra from the model estimate of peak acceleration (Johnson and Traubenik (1978)) are not tested and do not take into account high frequencies (20 Hz or more).
Id. at 35.
130. The Staff testifed that the same was true of the Regulatory Guide 1.60 response spectra which is derived from a composite of strong motion records recorded mostly from earthquakes greater than M = 6 and L
at different distance out to 100 kilometers. The Staff took the position that the Regulatory Guide 1.60 spectrun is less appropriate for the specialized problem of earthquakes in the magnitude 4 to 5 range at distances less than 10 kilometers.
Id. at 34. The Board shares this view.
1
- (ii) Response Spectra 131. The Applicant defined the ground motion from the maximum riagnitude 4.5 earthquake using the Hanks-McGuire model and assuming a hypocentral distance of 2 kilometers and an rms stress drop of 25 bars.
The resulting peak acceleration of 0.22g was converted to a response spectrum using the 84th percentile (mean plus one sigma) amplification ratios of Traubenik and Johnson. The Staff considered this spectrum appropriate for use in evaluating the effects of a reservoir-induced M
=
g 4.5 earthquake upon the plant.
Id_. at 35.
For structures founded on soil, the Staff took the position that the development of input spectra based on the application of this RIS spectrum at the rock soil interface would be conservative.
Id.
132. The Staff examined the sensitivity of its changed assumption regarding the depth of such an event and the 50 bar stress drop it recommended upon the Applicant's estimate of 0.229 and found that the effects of the increase in stress drop upon the estimated peak ground dCCeleration was more than compensated for by the effect of the increase in distance.
It, therefore, concluded that the response spectra derived from the Applicant's use of its model to describe ground motion from the maximum magnitude 4.5 reservoir-induced earthquake is conservative.
Id.
at 35. The Board finds this position both reasonable and convincing ar.d supported by the weight of the evidence.
(iii) Marrmoth Lakes Site-Specific Data 133. The Staff testified that it recently became aware of the availability of an extensive set of strong motion records recorded at an j
earthquake sequence near Mammoth Lakes, California in 1980. Thousands of records from over a thousand earthquakes in the magnitude 1 to 6 range
- were recorded. The Staff asked the Applicant to evaluate the data set so as to determine whether site specific spectra suitable for use in determining the ground motion from a M = 4.5 earthquake could be g
estimated. M. at 36; Reiter, Tr. 5907-08.
134. Based on its current knowledge, the Staff testified that this data represents the best source of data available to determine ground motion from an N = 4.5 earthquake in the 5 to 16 kilometer depth range.
L Id., Sobel, Tr. 5765. The average hypocentral distance of 7.3 kilometers associated with the data set used by the Applicant indicate that the resulting spectra would be a conservative estimate of ground motion within this range.
135. The Staff acknowledged that questions arise with respect to the use of these data in estimating ground motion at an eastern site.
These questions relate to reninnal differences in source characteristics and attenuation and differences in the site conditions at the Mammoth Lakes recording station and those at the Summer plant. The Staff did not believe, however, that these difference preclude use of the Mammoth Lakes data at the Sumer plant.
Id. at 37-38. The Staff stressed that the primary difference between eastern and western U.S. is that ground motion from eastern U.S. earthquakes is larger at greater distances. Reiter, Tr. 5899.
136. The Applicant's analysis demonstrated that at frequencies less than 7 Hz the Applicant's model spectrum would exceed the Mammoth Lakes spectra for all frequencies.
Id. at 38; see figure 1.
At frequencies greater than 7 Hz the Mammoth Lakes 84th percentile was approximately equal to the model spectrum except for a sharp exceedance centered about 8 to 9 Hz and a slight exceedance at 15 to 20 Hz. M.
137. The Staff regarded these peaks as consistent with those observed in the individual spectra and believed the peaks reflect the peculiar site conditions at the particular recording station. The Staff took the position that the !!amoth Lakes site-specific spectrum verified the conservatism of the Applicant's model RIS spectrum for describing ground motion from an M = 4.5 earthquake for those structures at the L
Summer plant founded on rock.
Id.; Sobel, Tr. 5767. The Board finds this position both reasonable and convincing and supported by the weight of the evidence.
(iv) Peak Acceleration Estimates for M, = 4.5 Earthquake 138. As an additional measure of comparison, the Staff compared the peak acceleration of 0.229 pronosed by the Applicant's use of the Hanks-McGuire model to anchor the response spectra describing ground motion from the M = 4.5 reservoir induced earthquake to estimated peak L
accelerations for an M = 4.5 earthquake derived most recently by other L
investigators using other techniques. M.at38;Reiter,Tr. 5907-08.
This comparison with recent estimates of peak acceleration versus magnitude and distance for different locations around the world indicated a wide variation in estimates with the Appiicant's assumed peak of 0.22g exceeding almost all of these estimates including those deemed most appropriate.
_Id. at 41, c.
Ground Motion Estimates: Shallow RIS 137. The Staff evaluated the sensitivity of ground motion estimates for its best estimate for a maximum shallow reservoir-induced earthquake (M = 3.0) against an assumption of larger (M = 4.0) earthquakes L
L occurring at shallow depths. M. at 41; Reiter, Tr. 5883, 5917-18, 5948,
- 5943. According to the Staff, use of the Applicant's model to determine ground motion from shallow earthquakes at close distances is not as appropriate as with normal depth earthquakes. As it earlier testified, the Johnson and Traubenik amplifications would not be appropriate where significant ground motion is expected at high frequencies such as have already been observed from nearby earthquakes at Monticello. M. at 41-42; Sobel, Tr. 5912.
In addition, the issue of saturation of ground motion with distance would have to be resolved if earthquakes at very close distances (within two source radii) were considered. M.at41-42.
140. According to the Staff, it is difficult to resolve this problem particularly with respect to the Brune model which by definition is a far-field model. The conservative use of the Hanks-McGuire model would be to estimate peak accelerations assuming no saturation.
Utilizing the maximum earthquake expected at shallow depths anywhere in the U.S. (M = 4.0) at its estimated shallowest depth of 2.3 kilometers L
(per Dr. Nuttli's testimony) results in predicted peak acceleration of 0.27.
Id. at 42. This estimate was made by the Applicant assuming an 9
upper fu of 30 Hz.
Increasing this to 40 Hz is expected to increase this estimate somewhat but, according to the Staff, it would still be less than 0.359, that single highest peak thus far observed at Monticello associated with the October 16, 1979 event. M. at 42.
141. That peak acceleration was recorded at a hypocental distance of 0.8 kilometers from a magnitude 2.8 earthquake at a depth of 70 meters. The Staff examined the effects an increase in magnitude would have upon ground motion estimates. Utilizing peak acceleration (that paraneter most related to spectral response at high frequencies), the
- Staff stated that one can estinate the ground motion at these frequencies. Assuming the scaling with magnitude reconinended by Joyner and Fletcher and a typical sc911ng with respect to distance, it was indicated that 0.359 for an N = 2.8 earthquake at 0.8 kilometers would L
scale to the same value or less assuming an M = 4.0 earthquake at a L
distance of 2.0 kilometers or more.
,Id. at 43.
142.
If a larger event were to occur, it would be deeper and the resultant ground motion from this event would be enveloped by the envelope of existing ground motions and estimated RIS spectra. Sobel, Tr. 5913. The Staff found 2 kilometers to be a very conservative estimate of hypocentral distance based on the distance of the plant to the earthquake clusters and Dr. Nuttli's estinate of 2.3 kiloneters as the shallowest depth at which an M = 4.0 event would occur. M.at43.
L Thus, the larger but deeper events would have acceleration values that are the same or less. Sobel, Tr. 5770.
143. The Staff emphasized that the purpose of scaling peak acceleration was not to arrive at a definitive estimate of peak acceleration from an N = 4.0 at 2 kilometers. Rather, the purpose was L
to provide some reasonable estimate as to the relative difference at high frequencies between the highest ground motion recorded so far at the Monticello dam abutment and the ground motion that may be recorded near the Summer plant from a postulated larger earthquake within the shallow zone of reservoir-induced seismicity.
Id. at 43.
d 144. The Staff concluded that the appropriate designation of the largest ground motion at the Summer site from the occurrence of RIS within the shallow zone of seismicity is the envelope of the response
. spectra fron data that have been recorded at Monticello.
Id. at 44; see figure 1.
This position is based upon the following:
(a) it is a very conservative description of ground motion from the Staff's best estimate of the maximum earthquake within the shallow zone of seismicity (ML" 3.0) and (b) utilizing both the Hanks-McGuire model and direct scaling of the highest ground motion recorded thus far gives reasonable assurance that if the largest estimated earthquake (M = 4.0) to occur in the L
eastern U.S. at a very shallow depth were to occur at the same depth in the shallow zone of RIS, the peak acceleration (and related high frequency motion) would not exceed the peak acceleration already recorded at Monticello (0.35 ).
H. The Board finds this position both 9
reasonable and conservative. This ground motion envelope exceeds the Applicant's proposed RIS spectrum at frequencies greater than 10 Hz.
d.
Status of October 20, 1981 Board Notification (i) Accelerometer Records 145.
On October 20, 1981, the Staff notified the Board that unprocessed accelerometer data from Monticello was recently made available by the USGS and contained a record whose peak accelerations and response spectra at high freque'1cies were greater than those previously recorded near the reservoir. Also the Applicant planned to conduct field tests to evaluate its hypothesis that ground motion record at the USGS accelerometer on the dam abutment may be amplified due to topography or other factors. The processed accelerogram data regarding this event was received by the Staff on October 19, 1981 in the form of USGS file report 81-1214 and forwarded to Board. M.at58-59.
' 146. The Applicant's evaluation of this data is contained in a December 2,1981 response to a related Staff question.
In that report, the Applicant showed response spectra for six events of magnitudes 2.2 to 2.8.
Observed peak accelerations for these events range from 0.12g to 0.34g and hypocentral distances range from 0.31 to 1.1 kilometers. Many of these free-field ground surface response spectra exceed the design SSE response spectra at high frequencies (above about 10 Hz). The Staff evaluation of these exceedances in terms of design is discussed in paras.
152-158 supra. As a result of its overall review (outlined above), the Staff concluded that small shallow events should be considered in the review of the seismic design adequacy of the Summer plant.
It concluded, as stated above, that an appropriate estimate of ground motion at the surface should be derived by enveloping the spectra of events recorded at Monticello.
Id. at 60; see figure 1.
The Board agrees.
(ii)
Field Tests 147. The Staff testinony contained a brief description of the Applicant's field experiment, the results of which are contained in the Applicant's November 19, 1981 report (Applicant's exhibit 42). The Staff found that the Applicant was unable to demonstrate that significant amplification of ground motion occurs on the dam abutment relative to the other free-field sites. However, the Staff concluded that a large reduction of ground motion was observed in the auxiliary building base-ment which is founded on bedrock relative to the free-field sites which are underlain by soil.
Id. at 61.
148. Reductions of spectral velocity in the auxiliary building with respect to the dam aLutment varied from 90 to 60 percent between 5 to 40
- Hz. The Applicant attributed the reduction and ground motion observed in the auxiliary building to either (1) the amplification of ground motion in the soil layer overlying the rock or (2) a building foundation effect, namely, the scattering of elastic waves incident at the foundation of the auxiliary building.
The accelerometer in the hydroplant, also founded on rock, showed peak accelerations 40 percent less than the dam abutment thus qualitatively confirming the reductions observed in the auxiliary building. H. at 61.
149. The Staff reviewed the Applicant's report and found that, since there are no free-field data for rock sites near the plant, the effect of the overlying soil layers itself cannot be simply calculated.
While the Staff believes there is a reduction due to the difference in site conditions, and some reduction would be appropriate for use, it had difficulty quantifying the reduction due to the fact that building foundation effects and site geology effects were not separately deter-mined by the field tests. The Board shares this difficulty. Another factor that makes quantification difficult is that ground motion from blasting may have different signal properties than earthquake ground motion.
Id. at 62; Reiter, Tr. 5812-13.
150. The Staff, nonetheless, cited approvingly the work of other scientists and the testimony of then Staff consultant, Dr. Nathan Newmark, in the Diablo Canyon proceedings for the proposition that a reduction in high frequency ground motion transmitted to the rigid foundations of large structures has been reasonably established. M.
at 62. The Board finds that the evicence supports a conclusion that some qualitative reduction of the peak accelerometer readings is appropriate.
68 -
The Board finds itself unable to adopt any rigorous quantitative factor for that reduction, e.
Seismic Design 151.
In preparation for his direct testimony, Staff witness James Knight, visited the plant site and made an audit of the design calculations employed for the Summer plant at the office of the architect-engineer, Gilbert Associates. The purpose of the audit was to ensure that the techniques used in modeling of the structures, the development of the response spectra for the evaluation of equipment and, in general, the design procedures, were fully competent. He concluded, based on that audit, that he could find no reason not to accept the procedures and results.
Knight, Tr. 5785.
152. Mr. Knight also went to the reactor site for the specific purpose of viewing the equipment at lower levels of the plant and to compare what he had seen at the Gilbert office (in terms of the layout and design) with the equipment types and categories in the nuclear plant and at the Fairfield Pumped Storage Facility. M. He noted the simi-larity in structure and equipment between the Summer plant and the hydroplant, particularly the electrical control equipment. M.
153. The witness noted that whatever ground motions actually took place during recorded microearthquakes were certainly felt by the Fairfield Pumped Storage Facility. The only effect there was that the unit tripped off line due to a vibration sensor. This would indicate that the plant did indeed receive some motion. This involved a so-called Mercoid or merccry switch which the NRC has not allowed in critical systems for years because thev are sensitive to small motions.
Id. The
Board finds this relative lack of damage to the hydroplant an important factor in assessing any safety hazard to the Summer facility from RIS events.
154. The Staff testimony on engineering matters went on to discuss the engineering significance of the fact that free field ground surface response spectra exceed the design SSE response spectra at frequencies above 10 Hz. The Staff acknowledged testimony arguing that ground motion effectively acting on the building foundation at the Summer plant will be significantly less than the free-field recorded motions. These reductions would result primarily from consideration of soil amplifica-tion effects that are present in the recorded motion and from reductions occurring due to wave incoherence at the structural foundations (therefore reducing the effective motion that is inpyt to the structure).
Id. at 63.
155. The Staff testified that experience in earthquake da, mage has demonstrated that a number of characteristics of seismic ground motion which have high significance from a seismological standpoint have little or no significance from an engineering standpoint. Principal among these characteristics is peak acceleration.
Id. at 10.
156. With regard to nuclear power plant structures, peak accelera-tions which occur, as they characteristically' do, as random high frequency spikes on the acceleration time history, do not represent a significant energy input to the structure. The response of the structures is essentially the same whether or not the peaks occur. The high frequency spikes do not contain sufficient energy to overcome the inertia of large structures and the frequency of the spikes is well above
. the response frequency of the power plant structures thus precluding resonant response. Jd.; Knight, Tr. 5771-72. The Board finds this testimony both reasonable and convincing and consistent with other material testinony on this subiect.
157. The Staff proceeded to examine the effect of spectral exceedances in question here on safety related structures.
It testified that these structures all have fundamental frequencies below 10 Hz, significantly removed from the peak high frequency motions characterized by the free-field response spectra.
Because of this difference in fre-quency, the response of the major structures to the high frequency motion will be low and less than the response spectra predicted by use of the SSE response spectra. According to the Staff, the stresses induced in the structures are controlled by the SSE response spectra.
Id. at 63; Knight, Tr. 5773.
158. The Staff next discussed the effect of spectral exceedanees on systems and equipment mounted in the structures. These will be excited by the notion of the structure at the mounting location (s) of the various components.
It noted that higher frequency motion (above 10 Hz) with little amplification will theoretically be present in excess of that predicted by the ground motion characterized by the SSE response spectra.
In the judgment of the Staff, prudence suggested that any evaluation include the high frequency motions.
Id. at 63-64. The Board agrees.
Systems and equipment mounted on or near the foundation slab will experience the high frequency motions directly transmitted through the slab. According to the Staff, it is for this group of components that l
I
- evaluation of the higher frequency notions are most significant. M. at 64 159. Systems and equipment with fundamental frequencies below approximately 10 Hz would by definition have less response than that predicted by use of the SSE spectra (since the SSE response spectra exceeds the free-field spectra below 10 Hz). Systems and equipment with fundamental frequencies above 10 Hz would experience a naximum response for the particular frequency shown on the free-field spectra. At approximately 25 Hz this response would be on the order of 1.5g; lower responses would be seen at other frequencies. At a response of 1.59 an object will undergo a force one and a half times its weight. Given the very short duration of the ground motion involved, an object would sense that force for an instant in each of the few cycles.
Id.
160. The Staff believed this to be an extreme example since it assumed the free field spectra to be operative in the building basement and assumed an object with a fundamental frequency at the peak response frequency of the free field spectra.
Id. Based on a review of equipment qualification programs for nuclear power plants over the past several years, the vast majority of systems and equipment employed in nuclear power plants have seismic capacities to failure well in excess of 29
.[d,.
161. The Staff pointed out an additional conservatism arising from the fact that the tests employed to explore equipment performance for qualification are very long compared to the ground motion under consideration here. The time history recorded on October 16, 1979 was less than one second in duration. Typical qualification tests last for
. approximately 20 seconds for each of five or more tests at maximum response.
Id.
162. The equipment and components now installed in the plant have been qualified by tests and analysis to the SSE design level. The practicalities of much of the testing are such that this equipment has, in fact, seen excitation at the higher frequencies up to perhaps 40 Hz.
iKnight, Tr. 5774. As far as structural integrity, much of this equipment can take very high peak acceleration loads. Knight, Tr. 5796.
163.
Further insight into the sensitivity of nuclear safety grade components to high frequency excitation (20 to 80 Hz and above) is avail-able through the extensive requalification testing being performed for Mark II and Mark III boiling water reactors. The firms supervising the test program report that inputs less than 60 Hz rarely cause n.alfunction and that where malfunction has occurred, the mode has been primarily minor contact chatter.
Id. at 65.
164. The Applicant has a present commitment to review all systems and components necessary for shutdown and continued heat removal to confirm that explicit margins exist for each safety component. M. at 5,
- 65. The Staff characterized this effort as confirmation that the equipment with high-frequency response on the lower levels of the facility have appropriate margins to perform their intended function for the life of the plant.
Knight, Tr. 5787. The Staff testified that reservoir-induced ground motion employed for this evaluation should, to the extent reasonable, take into consideration appropriate reductions in the free field spectra. H.; Knight, Tr. 5774. The Board agrees.
According to the Staff, greater relative risk could result from
- definition of excessive seismic input than from failure to consider the short duration high frequency motions associated with the reservoir induced seismicity. M.at65-66.
165. The Staff posited that factors of reduction on the order of 0.5 may well be appropriate for use with standard engineering design practices for assessing the significance of reservoir-induced seismicity.
Id. at 66; Knight, Tr. 5773. There is other testimony of record to support this view. However, it indicated that final definition of the reduction factor to be employed will require an integrated evaluation of site effects studies by the Applicant. The Staff believed that this effort properly formed a part of the confirmatory program underway. M.
at 66; Knight, Tr. 5773.
The Board agrees. The ACRS advised, and the Staff concurred, that this undertaking need not prohibit plant operation.
M.; Knight, Tr. 5774-75. The Staff believed this task could be completed during the early period of operation. Knight, Tr. 5787.
f.
Staff Pesponse to Board Consultants' Reports (a) Staff Evaluation: Joyner-Fletcher Report 166. The Staff notes that Dr. Joyner calculated the rms stress drop for the August 27, 1978 earthquake to be 32 and 25 bars (depending on which horizontal component is used). The highest rms stress drop currently proposed by the Applicant for any event is 29 bars for the October 16, 1979 earthquake. This is based upon an assumed reduction of 40 percent to account for the difference between the instrument site (dam abutment) and the rock foundation at the Summer site. The Staff took the position that the approximate unreduced stress drop of 50 bars is most r
appropriate for use in calculations involving the Applicant's ground
- motion model. Supplemental testimony at 67. The Board finds this appropriately conservative.
167. Based on the Applicant's observation that 98 percent of the seismicity is confined to the upper 2 kilometers, and assuming a stress drop invariant with magnitude, Drs. Joyner and Fletcher calculated a maximum magnitude of 4.6 for a vertical fault. The Staff considers the rigorous use of dimensions of seismicity patterns to derive maximum magnitude to be inappropriate where, as here, there is no evidence of a throughgoing fault. M.at68. The Board agrees. According to the Staff the difference in stress drop estimates for the small events recorded thus far and the difficulty in defining the maximum rupture dimension illustrate the uncertainties associated with estimating maximum magnitude from source paraneters. For this reason, the Staff did not place great weight on this technique to estimate maximum magnitude. H.
Nor do we.
168. Joyner and Fletcher note the results of work done by a colleague at the USGS, M. Zoback, who has interpreted stress measurements in the vicinity of the reservoir as indicating that the large stress drop events may be restricted to the upper few hundred meters. This would result in a maximum magnitude substantially less than 4.6.
A paper recently presented at a USGS workshop by Dr. Fletcher compared stress drops and Zoback's in-situ stress measurements and found that the data suggested that the larger events occuring at Monticello are releasing a large portion of the available stress. The observed dependence of stress drop on moment for shallow events in Monticello would then suggest an upper limit to the magnitude of these events.
In fact, Dr. Fletcher
appears to agree that frequency-nagnitude curves for Monticello events show an upper limit.
In their report, Joyner and Fletcher urged caution in placing major reliance upon this technique. The Staff has assigned a low priority to the return periods calculated for larger events.
Id.
at 69. See M. at 46-48. Joyner and Fletcher found that an examination of all cases of RIS in the Piedmont tectonic province is a more promising approach. M.at69. The Staff has done this.
169. Joyner and Fletcher proposed an alternative method for esti-mating peak ground motion from reservoir-induced earthquakes near the Summer site. They calculate peak acceleration and peak velocity for larger magnitude events by scaling the observed peak records for the August 27, 1978 M = 2.8 earthquake. They estimate peak accelerations of L
0.44, 0.62g, and 0.739 for magnitude 3.7, 4.3 and 4.6 events, respec-9 tively, at a source-to-site distance of 700 meters. The peak would occur at high frequencies.
170.
It is the Staff position that larger magnitude events would occur at greater hypocentral depths. A magnitude 4.5 event is assumed to occur at typical hypocentral depths observed in the eastern United States for that size event, namely, 5 to 16 kilometers. According to the Staff, modifications to the Joyner-Fletcher method to allow for increased source-to-site distances with increasing magnitude would significantly reduce the estinated peak accelerations and velocities. M.at69-70.
The Board agrees.
171. Joyner and Fletcher further indicate that their peak accelera-tion estimates are for high frequencies. The Staff does not regard it as appropriate to use high frequency peaks in conju'ction with amplification
- ratios determined from records processed in the normal manner which do not include high frequencies such as those recommended by Joyner and Fletcher or those used by the Applicant.
Id. at 70. The Board finds this Staff position reasonable and convincing.
172. Dr. Joyner suggested that their estinates of peak horizontal acceleration and velocity could be used by the Applicant to estimate spectral response values with the aid of appropriate amplification factors, which should be derived from an analysis of an appropriate collection of records digitized at adequate sample rates.
173.
In the judgment of the Staff, although this may be an interesting study, it is not necessary because the ground motion effects of the maximum shallow event in Monticello (which contains high frequency motion) can be modeled as the envelope (curve which exceeds at all frequencies) of the spectra calculated from observed Monticello records.
The Applicant's RIS spectra is appropriate for the deeper (and relatively lower frequency ground motion) magnitude 4.5 event. M.at70. The Board earlier noted its fundamental agreement with this position.
174.
In conclusion, the Staff found the Joyner-Fletcher review to contain implicit support of its methodology for deriving the maximum reservoir-induced earthquake.
It found their estimates of ;round motion at high frequencies inappropriate for use with standard amplification ratios and inappropriate in light of the Staff position about the depth of this earthquake and the maximum earthquake occurring in the shallow zone of seisimicity. H. The Board agrees.
4
_ 77 _
(ii) Staff Evaluation: Trifunac Report 175. The Staff testified that it knew of no regulation or precedent that would require the description of ground motion for the peak acceleration associated with the SSE to have only a 5 or 10 percent chance of being exceeded as recomended by Dr. Trifunac.
Id. at 71-72.
flor does the Board.
176. The Staff referred to the Applicant's evaluation of Dr. Trifunac's report where it was shown that a recent statistical analysis of peak acceleration versus intensity performed by Murphy and O'Brien resulted in significantly lower estimated mean accelerations for a range of intensities than those predicted by Dr. Trifunac in his peak acceleration model (Figure 1).
177. Dr. Trifunac's estimates of peak acceleration for magnitudes 5.0 and 5.5 are based upon a multiple regression which assumes that one can relate magnitude, distance, site conditions and direction of motion in one equation.
In addition, the basis for extrapolation of acceleration to "zero distance" is the attenuation curve developed by Richter for Southern California earthquakes from data recorded at distances greater than 20 kilometers. The Staff testified that it thought the direct approach which it utilized is less dependent on extrapolation and also is more appropriate for South Carolina where "zero distance" causative faults have not been defined.
Id. at 72-73. The Board agrees.
178. With regard to Dr. Trifunac's Uniform Risk Spectra (URS), the Staff regarded the seismic zones used in the acdel as not being justified by Dr. Trifunac. The Staff speculated that the distribution function of spectral amplitude as a function of intensity appeared to be based upon
- the sane data set as statistical assumptions used in Dr. Trifunac's peak acceleration model.
It advocated that these assumptions be examined in light of more recent and expanded analyses, such as Murphy and O'Brien.
.I d.
179. The Staff additionally pointed out that Dr. Trifunac's estinates make no allowance for spectral shape associated with a given intensity changing as a function of distance. The Staff testified that differences between the return periods calculated by Dr. Trifunac and the Applicant point out the uncertainties in probabilistic calculations for this site and are the reason the Staff assigned a low priority to the Applicant's probabilistic calculations for estimating return periods of
/
larger events.
It is the Staff position that its use of site-specific spectra is a more appropriate deterministic calculation for local intensity VII and demonstrates the adequacy of the existing design for eJ the Summer plant.
Id. at 74-75. This is consistent eith the weight of a
the testimony on this subject and the Board does not place much reliance
?
on such probabilistic approaches.
t (iii) Staff Evaluation: Luco Report 180. Dr. Luco evaluated the use of the Applicant's ground motion model to estimate peak accelerations. iie pointed out that, because of uncertainties with regard to input parameters, such as stress drop and fu that are needed for the Applicant's model, it was prudent not to rely exclusively on this approach but rather to use it in conjuction with other estimates based on recorded data and available correlations. The Staff agreed and, as it testified, it examined other approaches to verify 1
79 the Applicant's RIS spectra for an 11 = 4.5 event.
Id. at 75. The Board L
finds this effort satisfactory.
181.
Dr. Luco's other major point was that the procedure used by the Applicant to derive response spectral shape should be better documented.
In the opinion of the Staff, the Applicant, in its Additional Seismic Testimony, did this and compared its procedure (Johnson and Traubenik) with those found elsewhere. The Staff regarded the Johnson and Traubenik amplification ratios as the nost appropriate for an M = 4.5 event and has verified the conservatism of the L
Applicant's reccmmended response spectra through a direct comparison with site-specific data.
Id_. at 77. The Board agrees.
g.
Additional Comments of Dr. Carl A. Newton 182. Dr. Newton of the Los Alamos National Laboratory offered some additional comments regarding the establishment of the ground motion input to plant structures and equipment from reservoir-induced earthquakes. According to Dr. Newton, the results of the microearthquake monitoring program over the last year has increased his confidence that RIS will continue throughout the life of the reservoir as scattered microcarthquakes. The following elements influenced his position:
(a) only a few events have exceeded magnitude 2.0 and none have exceeded 2.5, (b) there has been no growth in the areal extent of the epicenters, (c) there has been no increase in the maximum focal depth and (d) except for a swarm of li = 1.5 events in April 1981, the frequency of events is at g
approximately the pre-impoundnent level.
He testified that these observations also add to the credibility of RIS physical models such n
- those utilized by the Applicant.
Id. at 79. The Board finds d
Dr. Newton's comments well taken.
183. Dr. Newton observed that the peak free surface acceleration from RIS thus far known is 0.35g horizontal acceleration (from October 16,1979 event). He posited several questions regarding this value. Does it reflect site amplification of the dam abutnent? Should it be the observed peak acceleration employed in scaling calculations for naximum RIS earthquakes? Should it be used to anchor a Regulatory Guide 1.60 design response spectrum? H.at80-81.
184 With regard to the first question, Dr. Newton testified that while there is no conclusive evidence that ground motion amplifications are produced by the dam abutment, time domain stress drops calculated from that accelerogram may be overcalculated due to the likely situation that the S-wave window used in that calculation included a surface wave arrival. M.at81. According to Dr. Newton, these waves would not be a problem to the structures. Tr. 5778.
185. He answered the second question in the negative.
Because of the possible S-wave contamination, it seened preferable to Dr. Newton to use the deepest event known to date when scaling for larger (hypothetical) RIS events. Supplemental testimony at 81.
186. With regard to the third question, Dr. Newton testified that, because the design response spectrum is a mean-plus-one-sigma spectrum and is used with a mean zero-period acceleration, he thought that in scaling motions to M = 4.0 and greater hypothetical RIS, the October 27, 1978 dam abutment accelerogram is appropriately conservative. Although it had not been done, he estimated that the August 27, 1978 spectrum
- scaled to higher magnitudes would envelope similar ones obtained from the October 27, 1978 accelerogram. M.at81-82.
187. Dr. Newton also performed some rough calculations where he derived " equivalent accelerations" for the three highest peak acceleration microearthquakes noted in his testimony. These were between 29 percent and 32 percent less than the observed peak accelerations.
Based on the notion that particle velocities are a more stable parameter than either displacement or acceleration, he indicated that his calculations suggested that the observed peak accelerations may yield overly conservative design response spectra. M.at83. The Board has taken Dr. Newton's additional views into eccount in arriving at its decision in the cise.
h.
Comments of Dr. Andrew Murphy 188. Dr. Andrew Murphy, whose views on certain matters differed from his NRC colleagues, appeared on the Staff panel. He did not participate in preparation of the Staff supplemental testimony and did not prepare any direct written testimony. Tr. 5781. He only had the opportunity to skim-read the Applicant's additional seismic testimony.
Tr. 578?.
Nonetheless, Dr. Murphy testified that his earlier position on the maximum reservoir-induced event possible at Sumer (ML = 5.0-5.3) had not been significantly altered. M.
189.
He adjudged the emphasis since the June,1981 hearing session as shifting to questions concerning strong ground motion, an area in which he did not consider himself an expert. Tr. 5779. He did not find the Staff position on the probable maximum earthquake that could occur in the upper 2 km beneath the reservoir (M = 3.0) adequately supported by g
the data. Tr. 5780. He did not offer an exact depth for an M = 5.0 -
L 5.3 earthquake stating it would probably be deeper than 2 km, but not necessarily deeper than the 5 km normal tectonic depth, Tr. 5780.
Dr. Murphy's views did not cause the Staff to alter its position on the issues. Sobel, Reiter, Tr. 5782-83. The Board finds Dr. Murphy's views about fault source dimension and depth (critical to his derivation of maximum magnitude estimates) to be against the weight of evidence. This, coupled with Dr. Murphy's limited analysis of pertinent testimony offered at this hearing session, compel the Board to accord Dr. Murphy's connents little weight.
III.
PRINCIPAL FINDINGS 190. The Board believes that the Staff has taken a thorough and conservative approach to the recognition and resolution of the important seismological and engineering issues presented in this case.
It finds the Staff's supplemental testimony reassuring and persuasive. Based on the testimony of record, the Board makes the following principal findings.
191. The best estimate of the maximum earthquake within the shallow zone of seismicity (less than 2 km) is M = 3.0.
See paras.
L 115-119, 140 supra.
192. The Board finds that the maximum reservoir-induced earthquake is an M = 4.5 event. These larger postulated events would occur at L
normal tectonic depth (5-15 km). See paras. 8-10, 71-75, 113-14 supra.
193. As a result of this magnitude-depth relationship, we are satisfied that ground motion acceleration would not increase as larger postulated RIS events are considered. See paras. 85, 140-142 supra.
L
194. The original RIS response spectrum proposed by the Applicant is appropriate for describing ground motion to be used in evaluating the effects of the maximum magnitude (M = 4.5) RIS earthquake upon those L
structures of the Summer plant founded on rock. For structures on soil, development of input spectra based on the application of this RIS spectrum at the rock soil interface would be conservative. See paras.
131, 137 supra.
195.
The maximum ground motion from smaller earthquakes in this shallow zone can be described by the envelope of maximum ground motions (respnnse spectra) already recorded at the lionticello reservoir. This spectrum is an appropriate description of ground motion at the surface for use in deriving input motions to structures founded on rock and soil.
This ground motion envelope exceeds the Applicant's proposed RIS spectrum at frequencies greater than 10 Hz.
See paras. 137-142 supra.
196. The Applicant has o Hered substantial arguments for the reduction of input motion at the structure foundation. The Board finds these arguments have substance but are not conclusive in a quantitative sense. We concur with the Staff that a quantitative assessment of specific reduction factors should be made as part of the equipment 1
l evaluation program in progress. See paras. 82-85, 145-148, 163 supra.
197. No one has suggested that these short, high-frequency motions have any significance with respect to structural integrity. The ground motions are occurring at frequencies considerably higher than the funda-1 mental frequencies of the major plant structures. There is limited, if any, observable damage from such motions in seismological and engineering l
evaluations of past seismic events and blast tests. See paras. 88, 90-92, 152-159 supra.
198. When the short, hiah-frequency motions are considered in comparison to the margins inherent in, and the design and testing processes employed for, safety-related systems and components (which may have higher response frequencies than the structums) there is reasonable assurance that the ability to shut down and cool the reactor would not be lost under the seismic shaking related to small reservoir-induced events.
Jd.; see also paras.15-19 supra.
199.
The fact that system and equipment responses resulting from normal plant operations often. exceed the maximum motions attributable to the reservoir-induced events at the Summer plant provides further positive assurance. There is considerable testimony that there has been no observed damage to engineered structures from events of less than M L
5.0.
See paras. 46, 57, 68, 88, 159-161 supra.
200. The Board finds that significant seismic design margins exist in the facility which provide Mequate assurance that the Sumer plant can safely cope with the largest expected reservoir-induced event at the site. See paras. 89, 91, 201. Despite recorded accelerations as high as 0.35g (from the October 16, 1979 event), there was no earthquake damage to either the Summer facility or the adjacent Fairfield Pumped Storage Facility. See paras. 88, 150-151 supra.
202. The Applicant has a present commitment to review all systems and equipment necessary for shutdown and continued heat removal to confirm that explicit margins exist for each ' aponent. The Board agrees
o.
with the ACRS and Staff that this activity need not prohibit operation of the Sumer plant. See para. 162 supra.
IV. CONCLUSIONS OF LAW 203. Based upon the entire evidentiary record of this proceeding, and upon the foregoing Findinos of Fact, the Board concludes that, contrary to the claims in Contention 4:
(a)Theseismicactivityinthe area of the plant site has been adequately described and evaluated by both the Applicant and NRC Staff with the result that the seismic design for the plant is adequately conservative; (b) site seismicity need not be monitored beyond the date considered reasonable by the NRC Staff.
V.
ORDER 204. WHEREFORE, in accordance with the Atomic Energy Act of 1954, as amended, and the Rules of Practice of the Comission, and based on the foregoing Findings of Fact and Conclusions of Law and this Board's Partial Initial Decision and Supplemental Partial Initial Decision, IT IS ORDERED THAT the Director of Nuclear Reactor Regulation is authorized to make such additional findings on uncontested issues as may be necessary to the issuance of a full-term operating license consistent with the terms of this Initial Decision.
205.
IT IS FURTHER ORDERED, in accordance with 10 CFR 962.760, 2.762, 2.764, 2.785, and 2.786, that this Initial Decision shall become effective and shall constitute, with respect to the matters covered herein, the final decision of the Comission 30 days after the date of issuance hereof, subject to any review pursuant to the above cited Rules of Practice.
Exceptions to this decision may be filed within ten (10) days after service of this Initial Decision. A brief in support of such
'l +
exceptionsmaybefiledwithinthirty(30)daysthereafter, forty (40) days in the case of the Staff. Within thirty (30) days after service of the brief of appellant, forty (40) days in the case of the Staff, any Other party may file a brief in support of, or in opposition to, such exceptions.
ATOMIC SAFETY AND LICENSING BOARD Herbert Grossman, Chairman Gustave A. Linenberger, Member Dr. Frank F. Hooper, Member Dated at Bethesda, Maryland this day of flarch,1982.
Respectfully submitted, l
b-Steven C. Goldberg Counsel for NRC Staff Dated at Bethesda, Maryland this 5th day of March, 1982.
I
UNITi'D STNIF.S OF &ZdICA NUCIIAR !!ECUIEIDRY CO&IISSION BEFORE Tile NIU11C SAFITY AND LICENSING BOARD In the Matter of
)
)
SOUlli CAROLIIM ELECI111C & CAS
)
COMPANY
)
Docket No. 50-395
)
(Virgil C. -Stunner Nuclear Station,
)
Unit 1)
)
CERTIFICA'IE OF SERVICE I hereby certify that copies of "NRC STAFF PROPOSED FINDINGS OF FACT AND CONCLUSIONS OF LAW IN THE FORM 0F AN INITIAL DECISION" in the above-captioned proceeding have been served on the following by deposit in the United States mail, first class, or, as indicated by an asterisk, by deposit in the Nuclear Regulatory Commission's internal mail system, this 5th day of March, 1982:
- llerbert Crossman, Esq., Chainnan Richard P. Wilson, Esq.
Adninistrative Judge Assistant Attorney General-Atanic Safety and Licensing Board S. C. Attomey General's Office U.S. Nuclear Regulatory Cannission P. O. Box 11549--
Washington, DC 20555 Columbia, South Carolina 29211 Dr. Frank F. Ilooper Joseph B. Knotts, Jr.
Administrative Judge Debevoise & Libennan j
School of Natural Resources 1200 Seventeenth Street, N.W.
University of Michigan k'ashington, DC 20036 Ann Arbor, Michigan 48109 Randolph R. Mahan, Esq.
[
41r. Gustave A. Linenberger S. C. Electric & Cas Company Administrative Judge P. O. Box 764 Atanic Safety and Licensing Board Columbia, SC 29218 U.S. Nuclear Regulatory Coninission Washington, DC 20555
- Atanic Safety and Licensing Board Panel i
George Fischer, Esq.
U.S. Nuclear Regulatory Carraission Vice President and Cencral Counsel h'ashington, DC 20555 1
South Carolina Electric and Gas Canpany l
P. O. liox 704
- Atanic Safety and Licensing Appeal Coltedaia, South Carolina 29202 Board Panel U.S. Nuclear Regulatory Caanissial l
Brett Allen '"2rsey h'ashington, DC 20555 Route 1, Box 93-C Little Mountain, South Carolina 29076
r
- Docketing and Service Section Office of the Secretary U.S. Nuclear Regulatory Cmmission Washington, DC 20555 l $.- A.Lik/
7 Steven C. Goldberg &
Counsel for NRC Staff 1
l I
l l
l l