ML19350B727

Rebuttal Testimony on Remand to Aslab Re Deterministic Methodologies on Earthquakes.Certificate of Svc Encl

ML19350B727
Person / Time
Site:	Seabrook
Issue date:	03/16/1981
From:	Chinnery M NEW ENGLAND COALITION ON NUCLEAR POLLUTION
To:
References
NUDOCS 8103230533
Download: ML19350B727 (22)

Text

-

l s '

( .- s

,N i v ,l'f'). # =xmo ss g .. r ,

% a! lggg y 12

' r, $$5hi g [

-> cc'r UNITED STATES OF AMERICA #

NUCLEAR REGULATORY COMMISSION /g @

BEFORE THE ATOMIC SAFETY AND LICENSING APPEAL BOARD 1

)

In the Matter of )

v'

)

PUBLIC SERVICE COMPANY OF ) Docket Nos. 50-44 ~

-)

~

NEW HAMPSHIRE, et al. 50 s (Seabrook St ati , nits 1 and 2) //

6- 120 &

REBUTTAL TESTIMONY OF DR. MICHAEL CHINNERY ,

S .J ON REMAND TO THE ATOMIC SAFETY AND LICENSING APPEAL D g\

SUBMITTED BY

~

THE NEW ENGLAND COALITION ON NUCLEAR POLLUTION Rebuttal Testimony The discussion in which we are presently involved illustrates the widely different requirements of the scientific method,~on the one hand, and the regulatory process on the other. ~The scientist most clearly realizes the imperfections of his data and his models, and is frequently forced to S

resort to statements of likelihood or probability, and the use of his professional judgment. The regulatory process is forced to make a concrete decision given all available information, sometimes acknowledging the fact that future data may indicate that a present decision was incorrect.

The basic problem is how to make the best use of scientific 0

T SS8 h 8103230533 /

advice in the formulation of this decision since, in order to be valid, the decision must be based on science.

The scientific method recognizes that no observation of a physical quantity is ever exact. In fact, a good scientist will never quote an observation without also quoting his best estimate of the error associated with his piece of data. In this sense, science and probability are inextricably linked together; a discussion of one inevitably leads.to a discussion of the other. The importance of this point cannot be' overestimated. Concrete statements of unavailable fact are seldom possible in science, and those that make them should be treated with suspicion. Similarly, an awareness of the inherent uncertainty in a result or a conclusion need not indicate a lack of technical ability, but often repre-sents a deeper understanding of the scientific problems involved.

A fundamental process in science is the attempt'to formulate natural laws or relationships from observed data.

We discover very quickly that when the data are sparse or i

i incomplete or in error by a substantial amount, a variety of laws can be proposed which are consistent with the observa-

,tions. Any attempt to select some of these laws as being more likely or more reasonable than others has to be based on the accumulated experience of scientists, or, in other words, a professional judgment. Over the centuries, we have developed one over-riding principle as an aid in approaching this problem. We always choose the simplest law that is

^- -

consistent with the data as a basis for further testing and prediction, unless we have a sound theoretical reason for seeking a more complex law. This principle (sometimes referred to as Occam's Razor) has been demonstrated to be the most productive way in which to approach the study of nature, and can be justified both on philosophical grounds and also using reccat results in information theory. .

The application of this principle in practice is not always easy. In particular, scientists may differ on the theoreticil reasons used as a basis for selecting a law which is more complex than the simplest one. The crucial point here is that the simplest law requires no justification, whereas any departure from the simplest law requires estab-lishing a sound theoretical basis.

The above comments are not trivial generalities; they have important applications in the present case. The his-torical data set of earthquakes in New England is both imprecise and sparse, particularly for larger events. More

_ geologic and tectonic data are available, but we can make few if any connections with the seismic data. We do not know the mechanism of earthquakes in this area. It is therefore hardly surprising that many different interpretations of the available data exist, and making a choice between these interpretations cannot be based on proof or disproof.

Instead, close attention must be paid to the proper. recognition of the uncertainties in all measured and derived quantities,

o and to the theoretical basis on which a complex interpretation is selected as being preferable to a simple interpretation.

As an example, the "Chinnery methodology" has been characterized as being "probabilistic," while the orthodox interpret:stion of Appendix A has been termed " deterministic."

~

The terms used are 1ess important than their implications; if " deterministic" is assumed to mean "without uncertainty,"

then I disagree most strongly. As Dr. Jackson has remarked

-s (Jackson testimony.p. 11), the Appendix A assumption that the largest earthquake that will occur in a province ~is to be found in the historical record for that province (regardless of the length of that record) implies the acceptance of a

~

level of uncertainty in itself. My biggest criticism of this assumption is that there is no scientific basis for the assumption that the largest historical earthquake is the largest that will occur in the province or that the probability of a larger earthquake is so low that it may be ignored. To the contrary, there is no way to evaluate the. magnitude of the uncertainty involved in the assumption. At least my approach offers a method for computing the risk involved.

In a similar way, if some law exists that relates fre-

,quency with intensity (and it is apparent from their testimony that Dr. Holt and Dr. Reiter feel that it does), then the simplest law that fits the. data is a linear relationship that extends over a much longer period than the length of the historical record. A law which is linear over the length of the historical record and then changes precipitously

is a very complex one, and hard to justify theoretically.

Similarly, the introduction of a non-linear relationship needs a sound theoretical basis, because it is not justified by the data alone.

In what follows, the "Chinnery methodology" is discussed in terms of the four assumptions mentioned in Dr. Reiter's testimony (p.3): 1. 4:triu; ,

1. Linearity of Frequency-Intensity Data (Reiter testimony, p. 3) "In a given region during a given period of time, there is a linear relationship between epicentral ~ intensity and frequency of occurrence. This re-lationship takes the form of logN = a-bI, where N is the cumulative number of events with intensities greater than or equal to I, a is a parameter which describes the level of activity, while b is the slope of the curve which describes the relative frequency of occurrence."

I prefer not to view this as an assumption, but rather as ius empirical observation. First, there does appear to be some sort of relationship between intensity and frequency, and second, the linear relationship is the simplest law Fiat appears to fit the data. This was found to be acceptable (with or without an upper bound cut-off) by 8 of the 10 respondents to question 3-1 in the Solicitation of Expert opinion (Tera Corp. Study,1979, Appendix 1 to Reiter testimony) .

If some more complex relationship is to be proposed, it most be justified on one of two bases. Either some sound thec etical basis for a non-linear relationship must be ex-pounded (none is available in this case), or the data must be unambiguessly shown to require a more complex curve.

1 4

While most scientists accept the linear relationship, a few 1

j have atte=pted to fit the data to a curve, with unconvincing 1 results.-1/

a .

1 l

Dr. Holt appears to prefer a non-linear relationship, but his arguments are not' convincing. First (Holt testimony,

p. 2), he has misunderstood the point discussed by Chinnery

I (1979), and states that addition of large events causes a -

I-non-linearity at the upper end of the frequency-intensity relationship. Chinnery (1979) calculated that the recurrence time of earthquakes in the Mississippi Valley for intensity X and greater was 537 years. However, on his Figure 4, Holt plots them as having a recurrence time of 56 years (because 7

three events occurred close together in space and time 170 22/ .

yearsago[7 of course this leaGs to a non-linearity.

By far the simplest and more likely explanation is that these earthquakes do, in fact, occur rarely, and that by chance we happen to have lived 170 years after the last -

large event in the Mississippi valley. There are many 1/ Dr. Reiter (testimony, p. 5) quotes three recent papers that have addressed this problem. Bloom and Erdmann (19FO) claim to have demonstrated a departure from the linear frequency-magnitude curve. In fact, when their quoted magnitude error of + 0.5 units is included, all of their regional data sets are consistent with a linecr .

. interpretation. Their global data set shows simply that there is a well known saturation in the magnitude scale at very large magnitudes. ~Berrill and Davies (1980) also are apparently unaware of the effect of saturation of the magnitude scale. Makjanic (1980) investigates ,

a data' set from Zagreb which includes intensities from I to VIII, and attempts to fit a very complex curve. If

~the data points frcm intensities I, II and III are omitted, the remaining points fit a linear relationship (sllpe about 0.52) very well.

-2/ It is very questionable whether that set of events should be regarded as three events or as a single occurrence.

I i

i

~ ^

other areas of the Eastern U.S. where large events have not occurred during the historical record, again by chance.

He also shows (in his Figures 5, 5A and 5B) some interesting data for the Charleston, South Carolina, region.

His frequency-intensity data for the whofe of South Carolina agree well with those quoted by Chinnery (1979) . However, data from the small area around Charleston have a radically different slope (0.25). The latter result is unusual and interesting. I note, however, that the area is very small (about 20Em~in radius), and that the seismicity here is likely to be dominated by the main shock and aftershocks of the 1886 event. Considerable further study would be necessary before concluding that this result is real, and that it has a useful application elsewhere in the Eastern U.S.

His arguments _about the significance of the WGC Earthquake Data Base are subject to substantial question. His Figure 7 (p. 17) contains a data point for intensity III earthquakes ,

over the period 1637 to 1979. However, we do not have a j

complete record of intensity III events since 1900, and data earlier than that.are meaningless. If this data point is i

omitted, his remaining data fit a linear model (with slope l

0.55) remarkably well (in fact, better than any data set I i

have seen for this area). The statistical significance of his quadratic model is marginal at best, and will disappear l

i if the intensity III point is removed.

Chiburis (manuscript) has analyzed this same data set with different results. He finds a remarkably linear i

l l

relationship for data in the range IV to IX for the period 1534-1977, for the northeastern U.S. and adjacent areas.

The slope for the uncorrected data set is 0.545, which agrees with Chinnery (1979). He then applies the method of Stepp (1973) to correct the data for completeness, and ,

concludes that his catalog must be missing 75% of all events at intensities V, VI and VII, while the data for intensity VIII are complete. The resulting corrected data have a pronounced change in slope at intensity VII. While the completeness of the catalog is always an important problem, the form of the correction seems to make little senso in this case. I find it hard to believe that 75% of'the events l

l with intensity VII are missing from the catalog. If, on the other hand, this is so, then the catalog may well be missing some intensity VIII and higher events.

My professional judgment is that each of the available

( studies that attempts to document a significant departure from linearity is suspect in one way or another. The errors in the data points (and in this context errors must include the effects of statistical fluctuations in rare events) do ,

not justify anything more complicated than a linear relation-l- , ship.

2. Uniformity of Slope (Reiter testimony, p. 3) "For those regions examined (New Madrid, Southeastern U.S. and New England) the data are consistent with other regions of both eastern and western U.S. In areas of little data an apparently uniform slope of 0.57 may be used to construct local frequency-intensity relationships."

l

p . _ . _ _ .

I

! i The question of the existence or non-existence of a

} regional variation in b-value for frequency-intensity rela-I tionships only has meaning once the linearity of the relation-

ships is accepted. Pr. Reiter (testimony, p. 5) stater that "the bulk of estimated b values lie between 0.4 and 0.6." I

! certainly agree with this. Many b values lie in tha range i

0.5 to 0.6 (see Chinnery,1979) . Some lower values (0.4 to 0.5) arise when the mavisus likelihcod =ethod is used for slope estimation.V This method is ideal if a atalog i

I

can be relied on to be complete, but leads to ancmalously I

low b-values if the data at lower intensit i es are incomplete, l as is the case here.

j Chinnery (1979) argues that the simplest approach to i these observations, certainly in the Eastern U.S., is that

] regional variations in slope appear to be small, and that, i

i given the errors and random variations to be expected in

! dais type of data, most data sets are consistent with a

?

' slope in the range 0.55 to 0.60. Notice that large slopes f

i give the smallest estimates of the annual risk from ?._ge

)

3 earthquakes, so this choice (while not in the middle of the observed range) would result in relatively low probabilities

} for high intensity earthquakes when compared to other pro-babilistic approaches using slopes in the range of 0.40 to 3/ This method involves giving greater weight to data points representing earthquakes most likely to occur, generally those in the range of intensities I-IV.

l t

i L

0.55. Accordingly, it would not be reasonable on the basis of available data to argue that the probabilities.shown by the slope of 0.57 are unduly high.

_. 3. Upper Bounds to Earthquake Size (Reiter testimony, p. 3) "There is no satisfactory way to establish the existence of an upper bound to the frequency intensity curve."

The existence or non-existence of an upper limit to the possible size of earthquakes in the New England area is a very important question, whether a probabilistic approach is taken to Part 100 Appendix A, or not. It is clear, from the Tera Corporation study (1979) that I take a particularly conservative point of view, and argue that, since we have no good scJ entific evidence tc: the contrary, we should assume l- -that intensities of X or more might occur. Let me refine

. -this argument again, in order to clarify it.

Dr. Reiter (testimony, p. 7) argues that geological, tectonic and stress information should all go into the estimate of the upper bound to earthquake size. I thoroughly agree that this should be done if it is available. However,

. such information is not available for New England. The only events that have been linked to a geological structure in

,this area are the 1940 events in New Hampshire, which occurred near the Ossippee Mountain rins dyke complex, and possibly some small events in the Connecticut River Valley, which follows a major structural boundary. All other events, both historical and instrumental, show no correlation with known geological or tectonic structures. In addition, we do 1

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

-n -

l not understand the causative mechanism for earthquakes in New England. As far as we know, the area is not subject to '

active tectonism, and we must therefore conclude that earth-quakes somehow are a result of overall compressive stresses.

(Sbar and Sykes, 1973) acting on ancient' tectonic structures.

Such a theory does not help us to estimate the largest earthquake that could occur in New England. ,

The one piece of information that we do have is a negative one. As Dr. Holt commented (teclimony, p. 4),

I there l's n~o evidence of faulting at the surface in New England. However, I question that this places any useful constraints on the problem. I will explain this with some i recent seismological data. .

.The absence of surface faulting constrains earthquakes in New England to be contained within the earth's midi and lower crust, This in turn places constraints on their fault dimensions (the area that actually fractures) . My estimate ,

is that.the largest fault area that might be constrained in 2

2 thi's way lies in the range 10 km to 100 km (the latter might correspond to a break with dimensions 10 x 10 km or 20 x 5 km). ,

In a recent paper, Liu and Kananori (1980) examined 5 mid-plate earthquakes (New England is a mid-plate region),

and their results are shown'in Figure'l. The fault dimen-sions of these events were found to lie in the range 10 to

.i 100 km,2 and the seismic moments were found to be in the i

1 e e ---

I 26 The corresponding stress range of 10 to 10 dyne-cm.

drops were found to be unusually high (in the range 100-1,000 bars; interplate earthquakes typically have stress 5/

drops in the range 10-100 bars).

In order to convert.the observed seismic moments into

? .

magnitude, Figure 2 shows a compilation of moment-magnitude data from Fitch et al. (1981). Clearly, a seismic moment in

' the range 1025-1026 dyne-en may correspond to a variety of magnitudes, depending on stress drop. In particular, a seismic' moment of 5x10 25 corresponds to an M s of 7.0 if the stress drop is 100 bars, and 7.5 if the stress drop is 1,000 bars. One has to conclude that mid-plate earthquakes appear to have higher stress drops than events at plate boundaries, and that, in spite of their small fault dimensions, their magnitude (Ms) may be in the range 7 to 7.5. This corresponds i

roughly to a maximum epicentral intensity of X.

-If magnitude 7 earthquakes can occur in mid-plate l regions. several questions remain. First, is New England a typical mid-plate region? In my opinion there is no sound l

, geological basis for saying that New England is in some way an unusual mid-plate region. Second, what traces of such I

, events (presuming they occurred in the mid or lower crust) i l f/

A seismic moment is a measure of earthquake size.

l l

5/ A stress drop is the difference between the stress existing before an earthquake occurs and the stress that remains afterwards.

l .

l l

would we expect to find at the surface? Intensity T events are expected to occur on the average every 6,000 years (Chinnery, 1979) in the Boston-New Hampshire zone. If this is so, only one may have occurred since the recession of the .

last glacial period, and in many parts of New England the .

overburden is sufficiently thin that it may be very difficult to recognize the traces of such an event if they exist at all. Intensity IX events occur somewhat more frequently .

5 according to Chinnery (1979), roughly every 1,500 years in the Bossod-New Hampshire zone. Similar arguments apply. ,

In my professional judgment, a magnitude 7 (M 3 ) earth-

. quake may well occur rarely in the Boston-New Hampshire zone, at a depth that may be as little as 5 to 10 km.

Furthermore, I feel it will be a long time before we get enough new information that we will be able to revise this estimate. As near as I can estimate,.a magnitude 7 earth-i

~

quake at a depth of 10 km would lead to a surface intensity

! of at least X. -

lI l

4 Extrapolation of the Frequency-Intensity Data (Reiter testimony, p. 3) " Frequency-intensity data may be extrapolated linearly to predict the probability of occurrence of earthquakes larger than those in the historical l

record. In particular, the probabilities of occurrence of intensity IX and X in New England may be predicted through linear extrapolation of the 160 years of historical data during which no such events occurred."

The previous assumptions lead inevitably to this assump-tion. If a linear frequency-intensity is a reasonable i assumption, and we are unable to place an upper bound to l .

l I

1 I

l ~ .: '

t .

)

intensity at less than X, then linear extrapolation is a logical step to take. .

It may be worth emphasizing one argument given in Chinnery (1979). The return period of intensity IX events in the Boston-New Hampshire zone was calculated to be about 1,400 years, and the return period of intensity X was found to be about 6,000 years. The probability that one event of intensity IX will occur during a 300 year historical record was found to be 19% (for an intensity X it was even smaller, 5%). So the absence of. events such as these in the historical record is completely consistent with the calculations.

Dr. Reiter (tistimony p. 9) raises questions about the validity of the data used in my analysis. The Smith catalog is indeed questionable in some details. To demonstrate that more recent data bases have little effect on my main conclu-sions, Figure 3 shows two other sets of data plotted on the same graph used in my testimony. The open circles contain data from a subset of an earthquake data base assembled by f

Tera Corporation and supplied to me on magnetic tape. The

_ area corresponds roughly to Southern New England, and the l data here are (I believe) taken from the work of Chiluris.

,Also shown, as crosses, are the data plotted by Dr. Holt in j the Applicant Testimony (Figure 7, p. 17). All three sets 1

l of data agree well.

Summary The definition of the Safe Shutdown Earthquake according to Appendix A requires an evaluation of the " maximum earthquake ii

' s

. 5

! h

.e

r potential." My professional judgment is that the maximum earthquake potential in the Boston-New Hampshire seismic }

.=

zone is at least a magnitude 7 (Ms) earthquake at a depth of

'f-perhaps 10km. Using the orthodox (deterministic) interpre- t L tation of Appendix A, such an event shouId be chosen as the ,

Safe Shutdown Earthquake.

Using my methodology, and extrapolating beyond the his- T h'

C torical record, a new piece of information becomes available, ,

-g and this is that such an event is likely to be very rare, {

recurring'on the average every 6,000 years or so. In this f

k case, the annual risk at the Seabrook site may well be sufficiently small over the lifetime of the plant structure n E

that it may be disregarded, and a smaller SSE may be chosen. .

{

In particular, we can estimate that the probability of occurrence of an Intensity IX event is roughly 10-3/ year in ff the province containing the site. A consideration of overall ba t

risk (perhaps 10-7/ year, as used by Farrar) and substantial {

safety factors may then lead to a choice of SSE' smaller than [

I' y

X. Clearly, the probabilistic approach provides a much more e' rational approach to the estimation of seismic risk. f

.?

. r h

s I

e h

t i.

E L

t r

s

~ .

.- . . . .. - J . m;* - ~ - -

I

-) s i -

i -

6 s g M* ,

~ N se _

• ~

1 (I -

60* '

.< A-

.. 3 PAome h 3( 4%

d 4

-.O*

O* . Sa l' t

• AMgRto l . r

- l Cu _ .

. . ausinAt ,

60*

r m, ^"l* MC ,; _ w.s '- '

, - 90*E 18 0

• 90*W O* 90*E *

• Fro.1. Clobal an'ap indicating tectonic plates and locations of earthquaken used in this study. The event r ombers are the same as lated in Table L Solid dote are thrust faults and open cirdee are normal faults.'-The - arrows show the directions of ---6= streen.

w .. _ .

3 i

- a I a s itj i a i # a t tl 6 6 61651 ,

Kanomori and -

~

_. Anderson (1975).

,10 s .

%, g.

s g'

N-

',r.

e ,.

~

60 '

u 800 bars e E .

• 2 .

- ~.

.v) -

~ '

g -

'I (El Golfo 1000

~

. _.s -- n

~

- 5**2 3e

,Truckee .

l ~ A e, g i I f I f r til I I I i e il f r i i r r r:

24 25 . 26 27 Log Mo. dyne cm -

Fac.10. Ier S versus log .tfe plot for the mid-plate events in this study (th- fault area platted here is j the average value for both fault planes) and the empirical relation forlarge eariquakes Of,ir,6).

t 1 .

l Figure 1: Evidence from Liu and Kar.a=ori (1980) that mid-plate -

! earthquakes (nu=bered 1-5 in the above diagra=s) =ay have s=all fault areas and large seismic cc=ents (in the range 1025 _ 1026 dyne-cm).

1 l

4 i

8 8 8

31 5 8 l

8 l l l 5

_ (.) -

30 s-

• INTER-PLATE (.) E_

o INTRA-PLATE

(.)

29 E- .

- - 5

=_ _

~

e *. g,

^

• 5 28 m =- -

.i e E ~

o

~

?

, _ ~ . . .

si a.el - .

c _ / o -

x j* .

" 27 =- - ..- -

=

o o 7 I

o -

o 7_

~

N. -

_- 2 ~/ o CP pO O O

_ .2 26 s- .

o

o. _

4 /. -

/ 7 oo o.

O . * *6

^

. _ 15 6*7 -

! 17e*f14 ' - -

25 s- '. <*9 5 .

• 12 *1 '~

13e _ .

1 .

~

5

! 24 5- ,

~

_ PICTON o l t l 1 l t

' l 23 7 8 9 -

~

4 5 6 l

Ms Figure 2: Data from a very recent paper (Fitch, el al., 1981) relating moment to magnitude. Moments in the range 102 5 to 1026 correspond to magnitudes in the range 5 5 to 7.0, if stress drops are about 100 bars. If stress drops may be up to 1000 bars, the M, value may be 7 5 or more.

. \

}

O I3 cSTo d de W N AM PS UtRE

\

8EfSNtc Zovs

\ S M 8TH 94rA: 17do - l 9 f 7

\\

• o s our- Aua ru un,o

, \\

• ff . I - If-3 , L e s & vv>< 70-74 o

L AT lTu3 E jo

\

-Tcraa CoAP. T A r-a G a-s e - .

\g 17o0 - FAes ar

.\ t/sw b e.srsite zoac

'\ x Behros F4 3 *rA- 8 *s e LJGC S Pe t.: c A*- "T~rste Mo a Y F t7_ .

~

15' - -

to

~

. ( GT

~ *

. pg

. ~ (yn l . .-

x\

\ _ _

I -2 \

10 - \ , 175 S caec Area cany;tavares -

loo

\o -

I i l'

\_

~

(o SER SRocw \

516 SSG '+ ,

\\

\\

~

-3 g _ ~

loca

. \

s \

~

\ \

\

\

\ \ .

~

\

,p . , , ,

i .

E E m_ vm- I zg- _ru r,,, , , y 3 4-5 6 7 twa an row e c:~. . . l *2

~~

-_.-....... L G n=

^ ~^

-. - - . - . . ~ .-. T . _.~. ----- > ~ -

References Tera Corporation, Seismic Hazard Analysis: Solicitation of Expert Opinion, Report to Lawrence Livermore Laboratory, August 23, 1979, NUREG/CR-1582, Vol. 3.

t Bloom, E.D. and Erdmann, R.C., The observation of a universal shape regularity in earthquake frequency-magnitude distributions, Bulletin Seismological

' Society of America, Vol. 70, pp. 349-362, 1980.

Berrill, J.B. and Davis, R.O., Maximum eutrophy and the magnitude distribution, Bulletin Seismological Society of America, Vol. 70, pp. 1823-1831, 1980.

i Chinnery, M. A., A comparison of the seismicity of three regions of the Eastern U.S., Bulletin Seis-l l

mological Society of America, Vol. 69, pp. 757-772, 1979 (Exhibit 2 to Chinnery Statement) .

l Stepp, J. C., Analysis of completeness of the earthquake sample in the Puget Sound area, in Contributions to Seismic Zoning, editor S. T. Harding, NOAA TR ERL 267-ESL 30, pp. 16-28, 1973. ,

Liu, H. and Kanamori, H., Determination of source para-meters of mid-plate earthquakes from the wave forms of body waves, Bulletin Seismological Society of i

l America, Vol. 70, pp. 1989-2004, 1980.

l L

^

v =.u . . : - . . . ..-...=:. . . . . . . - ~ . _ ..._. . _.

Fitch, T. J., North, R. G. and Shields, M. W., Focal .

depths and moment tensor representations of shallo" earthquakes associated with the Great Sumba earth-quake, Journal oj Geophysical Research, 1981 (in press).

Sbar, M. L., and Sykes, L. R., Contemporary compressive stress and seismicity in Eastern North America: an

,_ example of intra-plate tectonics, Bulletin Geologi-cal. Society oj America, Vol. 84, pp. 1861-1882, 1973.

Makjanic, B., On the frequency distribution of earth-quake magnitude and intensity, Bulletin Seismologi-cal Society of America, Vol. 70, pp. 2253-2260, 1980.

0

. . . - ~ .. , - , _

- = . _ _ - . -.: ; .. . . - - --.. .: . . - -. .

-y .. i_..c- 'w.OO7' - D. .* L ._ . . . - # -_ - - - . . - - - '

UNITED STATES OF AMERICA NUCLEAR REGULATORY COMMISSION BEFORE THE ATOMIC SAFETY AND LICENSING APPEAL BOARD

) ~

In the Matter of )

)

, PUBLIC SERVICE COMPANY OF ) Docket Nos. 50-443 NEW KAMPSHIRE, _e t _al. ) 50-444 (Seabrook Station, Units 1 )

and 2) )

)

.~

CERTIFICATE OF SERVICE

'I hereby certify that copies of the " Rebuttal Testimony of Dr. Michati Chinnery on Remand to the Atomic Safety and Licensing Appeal Board Submitted by The New England Coali-tion on Nuclear Pollution"have been served to the following parties on this 16th day of March, 1981.

l Alan S. Rosenthal, Chairman Dr. John H. Buck Atomic Safety & Licensing Atomic Safety & Licensing Appeal Board Appeal Board U.S. Nuclear Regulatory U.S. Nuclear Regulatory Commission Commission Washington, D.C. 20555 l

Washington, D.C. 20555

! Frank Wright, Esquire Assistant Attorney General Assistant Attorney General Environmental Protection Division Environmental Protection- Office of the Attorney General Division ' State House Annex, Room 208 Office cf the Attorney General Concord, New Hampshire 03301 One Ashberton Place ,

Boston, Massachusett0 02108 Thomas G. Dignan, Jr., Esquire Ropes & Gray Robert A. Backus, Esquire 225 Franklin Street .

O'Neill, Backus, Spielran, &c Little Boston, Massachusetts 02210

• 116 Lowell Street Manchester, New Hampshire 03101 Docketing and Service Section U.S. Nuclear Regulatory Commission Roy Lessy, Esquire Washington, D.C. 20555 Office of Executive Legal Director i U.S. Nuclear Regulatory Commission i . Washi ngton, D.C. 20555 _.

-_ .. =

.;.. [ ;

. . -_. . _7 . .

.- 2 --

Dr. W. Reed Johnson D. Pierre G. Cameron, Jr., Esq.

Atomic Safety & Licensing General Counsel Appeal Board Public Se'rvice Company of U.S. Nuclear Regulatory Commission New Hampshire Washington, D.C.' 20555 1000 Elm Street Manchester, NH 03105 Ms. Elizabeth H. Weinhold 3 Godfrey Avenue Atomic Safety & Licensing Hampton,'New Hampshire 03842 Board Panel U.S. Nuclear Regulatory ..

. . . . . - , e : , . .. .

. .. _ ,.. Commis sion -....__- .

Washington', D.C. ~20555 a .r.; %:{d:qn. y . ',i's . ,

9.r,- -e .. - -. , , .,. . q:l -

,.c .a - - ., . .. . .

? $,';*

, .  ? N.- ..~ L~ . J l '.- 1 i?..f

~

.~~~..

3 .t d5W ?hn William S FJordan', III~

I - .

,.,z

• ..~;  :. ,

• s, .,,%* p?, ,b .

• /

-w. ..

,* **  ;,. [

• .r' '4

, s

'.'-y * ,. -

J r . 1 ,

d

,[ E*

M'

~

'y.

. , - k. s',y

{.

a q,

r . .,,?_

s'.E' ' . .,

• p ;, l

. ~ ,_

• a, '

.e.,f

. s

• d e

4 s.

8.

,o e _y -~ , . - - - - - - - - -

- -.. -- ., , .- - ,