ML19345F961

On Application of Probabilistic Method to Estimation of Seismic Risk at Seabrook Nuclear Power Plant Site

ML19345F961
Person / Time
Site:	Seabrook
Issue date:	02/28/1981
From:	Chinnery M NEW ENGLAND COALITION ON NUCLEAR POLLUTION
To:
References
NUDOCS 8102190612
Download: ML19345F961 (52)

Text

{{#Wiki_filter:. UNITED STATES OF AMERICA NUCLEAR REGULATORY COMMISSION BEFORE TIIE ATOMIC SAFETY AND LICENSING APPEAL BOARD ) In the Matter of ) ) PUBLIC SERVICE COMuANY OF ) Docket Nos. 50-443 NEW HAMPSHIRE, et J.1. ) 50-444 (Seabrcok Station,." nits 1 and 2) ) ) STATEMENT OF DR. MICHAEL CHINNERY ON REMAND TO THE ATOMIC SAFETY AND LICENSING APPEAL BOARD SUBMITTED BY THE NEW ENGLAND COALITION ON NUCLEAR POLLUTION t I m m s / ,oc e nt i k t.sw " .fi + 2 FEB 17199t > E Ottice of the Secretwy } \\ Oxnet.rt i Iemce g Stucs / rc y { D50 3 s THIS DOCUMENT CONTAINS f/ POOR QUALITY PAGES Olo4* ig1 0219 i-CD

ON THE APPLICATION OF THE "PROBABILISTIC" METHOD TO THE ESTIMATION OF SEISMIC RISK AT THE SEABROOK NUCLEAR POWER PLANT SITE by Michael A. Chinnery Leader, Seismology Group Lincoln Laboratcry, M.I.T. 42 Carleton Street Cambridge, MA 02142 Phone: 617/253-7852 l ' l February 1981 l i l l

J 1 i ~ j Introduction l 1 i The Safe Shutdown Earthquake (SSE) is defined as "that

I earthquake which is based upon an evaluation of the maximum l

earthquake potential considering the regional and local j geology and seismology and specific characteristics of local J l subsurface material" (NRC Rules and Regulations, Part 100, Appendix A, Section IIIc). There are two methods which have been proposed for the estimation of the SSE: (i) The " Deterministic" Method: In this case, the j largest earthquake in the historical record in I the tectonic province containing the site is ], taken to be the SSE. Some additional conserva-l tism may be included by making the SSE larger i than the largest historical earthquake, though 4 this has to be based on geological evidence. In l the Eastern U.S. the l'ack of detailed correlation { between seismicity and geological structure makes it very difficult to estimate the validity and amount of this additional conservatism. (ii) The "Probabilistic" Method: Here the historical i record is-taken as only a sample of the long term seismicity of the tectonic province, and an attempt t is made to extrapolate this relatively short record to longer time intervals. In this case, the con-i cept of the " maximum earthquake potential".used~in

,1 ] 4 the definition of the SSE has to be modified, and the SSE must be defined as that earthquake which will occur in the tectonic province containing the site with some fixed acceptable level of annual l risk or probability. This acce'ptable level of l risk is net defined in the NRC Rules and Regula-tions. The Nuclear 'Aegulatory Commission has ruled (Order CLI 33, 25 September 1980) that the second approach is not incon-sistent with Appendix A, given our present understanding of earthquake science. In what follows, we explore the applica-tion of this apprcach to the Seabrook site. The Historical Record j' In New England the historical record of earthquake occurrence is approximately 300 years long. The only catalog of seismic events in this area that.has been published.in the scientific literature is that , by Smith (1962, 1966). The earlier parts of this record are not very reliable. Instrumental records, again of variable quality, are avail-able since the 1920's, but only in the last few years has a proper seismic network been installed. This network has detected..elatively few events since it was created, and can-contribute little to the assessment of seismic risk in the area. Ne are, therefore, forced to work with the historical i data set,.in spite of its inadequacies. Now we have to ask two important questions: Supposing that we thoroughly

1 understood the long term seismic characteristics of the area, how well can we predict the seismic activity during the next 50 years (the anticipated lifetime of the Seabrook plant) ? And, is the 300 year historical record really representative of the long term seismic characteristics? Both of these questions are difficult to answer. The first is most easily disposed of, since if we cannot use the past to predict the future, we have to give up any attempt to estimate seismic risk. We assume at this point that a thorough characterization of the seismicity in the past is indeed a reasonable basis on which to ecmpute future seismic risk. The second qu tion cannot be disposed of so easily, and lies at the heart of all controversy concerning the i' estimation of seismic risk. How can we'use the historical record to make the most reasonable estimate of the long term seismic characteristics? In ord,er to tackle this question, it is convenient to consider the spatial distribution of earthquakes separately from the distribution in time and size. These two aspects are discussed in the following sections. i Selecticn of a Tectonic province The concept of a tectonic province is a legal one (a s - defined in Appendix A), und has no clear scientific signifi-cance. The proposition that earthquakes must in some way be

alated *e neological structure and tectonics is inescapable, but it is not at all clear that large provinces _can be defined'

.-. _ = _. _ -.... i 1 i -4 l within which the seismo-tectonic characteristics are in any I sense uniform. Attempts to define such provinces asually lead {i to a wide range of interpretations (see, for example, McGuire j l 1977 and Tera Corp. Study, 1979). These difficulties certainly I i i apply in the case of New England, Figure 1 shows a map of the epicenters of earthquakes listed in the Smith (1962, 1966) catalog. Marked clusters of events near the Seabrook site occur in Southern New I Hamoshire and in Northeastern Massachusetts. In previous I studies (Chinnery and Rodgers 1973, Attached as Exhibit 1, and Chinnery 1979, Attached as Exhibit 2), these two clusters have been included in one seismic zone (or tectonic province), i 4 and this is indicated by the broken line in Figure 1. In what follows, we use this Boston-New Hampshire i seismic zone as the tectonic province eppropriate to the t j Seabrook site, recognizing that only weak arguments can be made for any choice of tectonic province in this region. (Tera Corp. Study, McGuire 1979) The present choice is at-least a reasonable one for the historical local seismicity, since the population density has been highest in this parti-l cular area. Instrumental epicenters for 1975-79 (see - Figure 2) are roughly consistent with this choice; the cluster of epicenters in Southern New Hampshire can still be seen,.but recent seismicity near Cape Ann, Massachusetts, . in apparent contradiction to,the historical ( has been low record). Certainly, neither the historical' record nor the i instrumental record lead to any good arguments for isolating 1 .m r

- ~.. .r- ...~ _3_ _ N }. k9 kb.- i' / l/ \\ /. s O c/- o eg eN, 3 o O Q b _, \\\\ o a o (

7. O a

A ol x@ O f 'O 0% O Q % =5 t c cD \\/ l e C 3 me ree e b Os \\ O g \\ oi

• Q iNTENseiv uaGMefuoE g

g 9 f, D I ') =- i h\\ n" gt \\- .I 7 l '~ ,6 g \\ l ct,. y U~ h f .s-i [ eoa'a. n 1 -a { j _O o } n g'- s ~=\\ 4,, n g* I .-r # ,,q.a.1,$ 1 ,P il- + o = \\ g Co , = ~(( 1 0 0 O =-- le $aa b' 1_._~~ h g g[ sij- -2 a .,y o \\ is e CNTNrRC (UneMeeHest 2 2cI) o o r acc h?st tuaces..hes y-201 carive twictura.ts o \\ C2 stiswocne-statiou i \\ .\\ \\ \\ \\ scatt m witts Fig. 1: Epicenters in "aw England fren the catalog by Sc:ith (1952, 1966). Tne broken lina indicates the Ecston-New Fanp:hira seis.:ic =cne (f. c Chinnery, 19'T 9a ).

p e o-o o o o o .o ~ e e e to a o c> a a v v v v ? T to (O \\g) N '. s ) \\ \\ \\ \\ v-g s(PN \\ )\\ .c m e m._ s 5i 21 \\" co 1 \\- e t0 e 2 T \\ .w os

• \\. %s i.M i t-e

\\ s. I"E . l, '\\ \\. e r .\\ 4 oae r a N / ), %,, ~ \\ =,c y s-N/' \\ -c.\\ \\ s s. .A n .. \\ ^. \\,$ d,. \\\\ $5c b \\ --:',s \\ ( 4,- " 35 ,~ \\ \\s v-Etd

• 8:?;, i t,J, r

\\ \\ \\. p ' 3 -a 1 l N s.~. \\., 'n ' c-

r

\\.

o.. -

s N !, %'s \\s s uoc Y \\s'yfM =*2 4. v 1\\ U I! L-( \\. i., .jz m\\ \\\\ <\\ \\ \\ \\ tai

n t

\\. \\ \\.wh l i B"E -u: 9oa

3 % u

\\ .e t \\.G'2 55 \\ 7 S? I i t-tu 4-I', $EE [ ~ 8 UD \\\\ K i se: \\ '\\d \\\\ d \\,

• h

\\ ) \\ -\\v.1 ' M n3.:. o=W / n. t / \\ \\ m -m s ,/ \\> ~-=

c I

\\ in S b d \\ 1. M \\ 8; \\ l \\ \\ \\bh -i...~., d3 \\ ) % \\ . /-r~ w e t

- the Seabrook site from seismicity,in Southern New Hampshire and Northeastern Massachusetts. In view cf the inadequacies of the historical record and the difficulty in selecting an appropriate tectonic province, assessment of the seismic risk at the Seabrook site can be based on a number of different assumptions. In my view, the most reasonable and most conservative assumption is that the seismicity of the Boston-New !.ampshire zone is a valid basis for estimating the risk at the Seabrook site. Frequency-Intensity Relationships The characterization of the seismicity of a province in terms of the rates of occurrence of earthquakes of different sizes is usually accomplished using frequency-magnitude or frequency-intensity relationships. In the present case we use the latter, since only intensities are quoted in the Smith cataleg. In addition, we use cumulative frequency-intensity

counts, i.e., we count the number of earthquakes larger than

^ or equal to a given intensity va1ue during a given period. The extraction of frequency-intensity data from a catalog such as Smith's must be carried out with care, since the completeness of the catalog at icwer intensities is likely to be a strong function of population density, and therefore of time. Ne use the approach described in Chinnery and Rodgers 1973 (Exhibit 1) here. Having extracted and plotted the data for the Boston-New Hampshire seismic zone, we have three important question to consider: i I f

. ~. 4 _g_ I (i) can the data be represen,ted by a linear frequency-intensity relationship? (ii) if so, what is the slope of the linear relation-ship? i (iii) is there some upper bound to th.e intensity of earth-quakes that can be expected in this seismic zone? Let us consider each of these in turn. 1. Linearity of frecuency-Intensity Data i Frequency-intensity data for the Boston-New Hampshire zone are shown in Figure 3 (taken from Chinnery 1979) (Exhibit 2). Clearly, the data are sparse. For the period 1300-1959 only six data points are obtained (for intensities II to VII) and-it seems likely that those for intensities II and III are un-f, reliable due to incompleteness (even though these points are I ( based on the very recent period 1928-1959). The remaining four data points actually lie in a relatively. good straight 2 i line, but the slope of this line (about 0.50) is, as we shall see below, unusually low, and wo'uld lead to high estimates for the rate of occurrence of large earthquakes. A more reasonable i.itercretation is that-the number of intensity.VII events (3) I during this period was unusually high, and that the intensity IV data set may be incomplete. 1 If these comments are valid, perhaps only the intensity-1 V and VI' data points are at all reliable, and'we can not make ~ h 1 any conclusions from the data taemselves about the linearity of'the frequency-ir. tensity relationship. ~ :Da this case, we must rely en information from elsewhere. ~ ~v ,-,~m eg,- e ,.g-m,-~+m -evr-w,-c--w,- -e., +--ww4,-*Ts'--- - =

---

N-'t'-*--T*r 1

• w-F r*-~~n-"'**--e d4

-~-~--'rT --t'***'r F-y

i. ~ t i

• 3T.**....

o. -( -

?- e ,s, O.5 BOSTCN-NiEW HAMPSHtBE . ~ 1800 -1959 e O i - * - ~ e s 3 1.sg N = 2,15 - C.59I c / ~ = 0.3 /- 0ll s 4 LJ 2 .?- Of i s.a Q- -1.0 ~ ~- 2 y u +l O. -a cm o _J -12 a = .-2.0 .i-e.. \\ q..- - e. I t f f I E H E 3 1 ' _' * A i INTU>IJ'rt >.. ~i..ev > "..v.- .- ", ' '. ~ -- \\. Fig. 3: Frequancy intensity data f: em the Easten-!!a'.4 Hangshir e seic=ie ene, derived frca the Saith catalog using tha cetheda siven in Chi.1nerf l and ?.cdgers (1973) (fec= Chinnery, 1979a). l --.-sev w-w swep ,.r..m e w-w,.mvm. .-ww-- v--.--ew4 -,,,y., ,-m,, ,gn- ,w-- <- --~ +

~ -- h -1o_ i. In my view, the current situation can be summarized as i follows: The vast majority of seismologists have accepted I the linearity of frequency-magnitude data as a working j hypothesis. (See, for example, Evernden 1970, Veneziano.1975, { and the references cited in those papersb. It is, however, still a hypothesis, with no clearly developed theoretical basis. And there are a few instances where non-linearities i are apparent in the data. These have led to several publica-tions proposing non-linear relationships, though in my view these can generally be attributed to poor or inadequate

data, i

The linearity of frequency-intensity data has been dis- {- cussed much less.. Several investigators have proposed 1 linear relationships between intensity and magnitude, (See, i for example,.Veneziano 1975) and, if these are valid, a linear frequency-magnitude relationship implies a linear 1 frequency-ictensity relationship. Of what scientific litera-ture there is, the vast bulk assumes that frequency-intensity. relationships.are linear (see, for example, references quoted in Chinnery.197.9) (Exhibit 2). One point'should be made here.. Intensity-(i.e., maximum epicentral intensity) is a very different scale from magni-tude, and the observed linearity in the relationship between the two at commonly observed intensities has no sound.theoret-ical basis. Certainly forevery large earthquakes there must be a departure from linearity, since-intensity has.an inherent upper bound.(intensity XII).while magnitude.is'an open-ended l i I L

i : i scale. Note, however, that all sc. ales become unreliable.for large events (roughly M>7), due to saturation and other effects. In summary, the apparent linearity of much frequency-intensity data must be treated as an empirical observation. 1 Its wide acceptance by seismologists suggests that it is-i ) useful as a working hypothesis. f 2. Slope of Frequency-Intensity Data 1 If we accept that in any given region we can expect a linear frequency-intensity relationship, the next question must be: Does the slope of this relationship vary signifi-j cantly from region to region? The only study that has addressed this point is Chinnery. l, 1979 (Exhibit 2). In that paper it was shown that there seems to be a remarkable uniformity'in the slopes determined from various areas of the Eastern U.S. Values of this slope l were typically found to lie in the range 0.54 to 0.60, and in fact, all the available data 'are consistent with a slope of 0.57. i l This is an important point for areas such as the Boston-t l New Hampshire seismic zone, where some of the data points I l I may be unreliable. If we assume that the data are to-be fit ith a straight 1.ne with slope about 0.57, then we can use I rhe most reliable data points (for intensities V and VI) to. j define the frequency-intensity relationship (see Figure 3). i In my view, more complex relationships are not justified by the data. i

3. Existence of an Upper Bound Intensity Having de:..ine, a frequency-intensity re.,ationship, we would like to use this to extrapolate beyond the historical data points, to give an estimate of long tern seismicity. The question remains: How far may we continue this extrapo-lation? Is there an upper limit to t.*e site of earthquakes that can occur in an area like the Boston-New Hampshire zone? If so, what is this limit? I have examined this question in some detail (see Chinnery 1979b, Attached as Exhibit 3). My conclusion is that we do not knew the answer to these questions at the present time. One aspect of the problem is worth mentioning here. All seismologists (including the author) agree that earthquake site (however measured) cannot increase in,e:1 nite,.y. Physical constraints arising from the earthquake source mechanism will set a limit to both so: tree dimensions and strain release. On a global scale, this upper bound is at a rather high level, somewhat above the lar:est kncwn earthc.uakes. On a regional level, l much less is known, and there is considerable disagreement 1 1 between the (guess-) estimates of different seismologists. 1 i 1 In a recent study (Tera Corp. S tudy 1979), ten experts I l in the seismicity of the Eastern U.S. nade estimates the .w a.-o .,a 1.. ...o w - 4 ~, t,. wa 3,.,- 3 o-.

2. e, ~4 - o... e,.---.4.,..e.,.4...

.3. a -: .a. .e-.. Cape Ann, assachusetts region. These are listed in Table 1, and Itlustrate the disagreement clearly. There is little l point in averaging opinions such as these. Notice, however, ..e

7. O, o.<._r aA..;.-

o.w.. n. o s a.4 w ; l.4.. .. a. w....e .w .wm. e .w .uo a v- .p y -u upper bound to earthquake size may be X or greater in this region. '1" Tw 1 O A Estimates of the Larcest Earthc.uake Expected to Occur in the Cape Ann, Massachusetts Region (Tera Corporation, 1979) Expert Low Estimate Best Estimate High Estimate 3 IX X XI 4 VI VI X Y. ' 'A . A 10 VII VIII IX 13 IX X XI 6.2 6.4 o.i i S 6.0 t .. 't 6.2 6."i Qs 1, 0.0 6.s-i.0 4 12 5.75 6.25 i (Here, arabic numerals indicate magnitudes; as a rough conversion to intensities, 6.0 VIII and 7.0 IX or X.) i i In mv view, the oniv. valid conservative. interpretation + l of this set of oc.inions is that we should admit the n.ossibility I, .of an intensity X earthc.uake in the Boston-New Hamo. shire i

c:

n...a ,..e a. aes-a.... v,ene,., _ -, ca .._:.3c i n. e. .e.m...---- m .c m s. ~1/ that will persuade us to revise this value. -1/ In it.= previous ruling, th-a Acpeal Board indicated a ci: culty in accepting t.nat da a trom one area coulc.,ce c any use in attempting to project the seismic charac-teristics of another area. This problem is fully t.- .2_.,_.. _... 23 . 3-

I 5 Estimation of Seismic Risk at the "Seabrook Site In the above sections we have laid out our basis for the evaluation of seismic risk at the Seabrook site. To summarize: We have selected a " tectonic province" contain-ing the site, which extends from Southern New Himpshire to Northeastern Massachusetts. Following Appendix A (section V, para.

a. l. ii), we assume that the largest earthquakes 1

that can occur in this province will occur at the site. Frequency-intensity data are extracted from Smith's (1962, 1966) catalog using only data after the year 1800. Through these data we will fit a linear frequency-intensity rela-tionship, with a slope of about 0.57, and use this as a basis for extrapolating to obtai ' measure of long term I seismicity. Extrapolation of the line is considered valid out.to an intersity of about X. r The result of applying these procedures is shown in Figure 4.. The data points are t,he same as shown in Figure 3. The. solid line has a slope of 0.57. Broken lines indicate slopes of 0.50 and 0.68; these would appear to be very wide bounds, based on other data from the Eastern U.S. (Chinnery 1979, Exhibit 2). The 1955 Cape Ann earthquake occurred a litte over 200 years ago, and has been estimated to have had an epicentral intensity of between. VII ~and VIII. 1 1/ discussed in Chinnery 1979 (Exhibit 2). The empirical observation was there presented that data from three areas of the Eastern U.S. are consistent with a uniform frequency-intensity slope of about 0.57, and that the data contained no evidence for the presence of a limit to earthquake size in these areas. This is an empirical observatic. and ts independen; of the geological char 22-Ceri3rics of the three areds. J

-n-f3osro4 //eW N A-M PS } Htls Seusm e zeaa \\ \\ G Arr A : I T o o ~ f 9 5~ 7 s\\, \\ \\\\ o o \\ -f Io - i \\ 1 ~ i 9\\ 1 6 \\.\\ 4 .. ti, ..g \\ ?-y_.'.ll 6 SEsT 3PM*R 4 -g to lo .t p. .5 ~ .r.. g_ : ~ :.-. p ~ J- . :M!* \\ ,i. \\g

-m

. c.. t. .. =.- \\ _~ \\ 175r Caes Asa E M ruqu;Hes 1 I-l o toc ~ ~ ...N

• j

\\

.u.-.-

r. 1 .g .."5.. ],,.y g Seesacew Li: SSc 7 o ;. e . y:.. =

. y..

'. k.;*, ?- .~~~ \\, .,x s l 1o - n==== t -, cc . S t.c e r a S 7 i l \\ \\. 9 ~ t \\7 a So l o 49 A \\ 1, \\. \\ l l 2. _= l

3

- t. 31.- _.k !- .... 2.iLL A. .....c.ii. , r c u s r: y L...- LL I Fig. 14: F.xtrapolatien of frequency-ir. tensity data for the Bosten-New L g

4 4 f 1 5 J i The open rectangle shows how this earthquake would plot on i l the present diagram. Clearly, that event is consistent with. 3 t our extrapolation from later data. The current Seabrook SSE of VIII is found to occur with 1 an annual risk of abcut 10-2.5 (this corresponds to a t -3 s return period of about 300 years). An annual risk of 10 4 (return period of 1000 years) corresponds to an intensity -4 IX, and an annual risk of 10 corresponds to an intensity j of at least X. The problem that remains :s to define the acceptable level of risk which will define the choice of the SSE. -3 -4 Though numbers in the range 10 to 10 per year have been i mentioned in the past, I am not aware of any formal definition of this risk, which clearly involves many societal, economic and political factors. I Conclusion This case study of the application of the "probabilis-i tic" method brings out all the main features of the method. Most important, it indicates that the definition of the Safe Shutdown Earthquake must be accompanied by a definition of the acceptable annual risk of the occurrence of the ground motion corresponding'to this size of earthquuxe. I 1 i

i a i i ?.. References

Chinnery, M.

A. and Rodgers, D. A., Earthquake statistics in l Southern New Englund, Esrthcuake Notes, vol. 44, 4' pp. 89-103, 1973 (Exhibit 3). i

Chinnery, M.

A., A comparison of the seismicity of three i regions of the Eastern U.S., Bulletin Seismological s 1 Society of America, vol. 69, pp. 757-772, 1979 (Exhibit 2). 4 i i

Chinnery, M.A., A study of maximum possible earthquakes, in Annual Report, NRC Contract NRC-04-77-019 (NRC Publica-I tion NUREG/CR-0563), 72p., 1979b.

(Exhibit 3 is the same study as publishad by Lincoln-Laboratory, MIT). i l' I l Evernden, J. F., Study of regional seismicity and-associated problems, Bulletin Seismological. Society of America, vol. 60, pp. 393-446, 1970. i f

McGuire, R.

F., Effects of uncertainry in seismicity on esti- \\ mates of seismic hazard for the East Coast of the United i States, Bulletin Seismolecical Society of America, vol. t l l 67, pp. 827-848, 1977. I

Smith, W..E.

T., Earthquakes of Eastern Canada and adjacent areas 1534-1927, Publications of the Dominion Observatory,. I Ottawa, vol. 26, pp. 271-301, 1962. I v

Smith, W.

E. T., Earthqu. .,s of Eastern Canada and adjacent areas, 1928-1959, Publications of the Dominion Observa-tory, Ottawa, vol. 32, pp. 37-121, 1966. Tera Corporation, Seismic hazard analysis: solicitation of expert opinion, Report to Lawrence Livermore Laboratory, August 23, 1979, NUREG/CR-15 8 2, Vol. 3. Veneziano, D., Probabilistic and statistical models for seis-mic risk analysis, Publication R75-34, M.I.T. Department of Civil Engineering, 1975. Qualifications An updated resume is attached as Exhibit 4. l 1 < l t r i

• I t

l

f .n 1 4 "L. 4 -..r. n.J q m. f t..r wq ar .r-t*_. r. 416 w n .nre. I VA w .U.,,, . n 7.. c.r. : L, e A u.. :r - -a p, r,... w...In,S10s . w: \\ 3: unr .n: n a v... r n Sn.

n.. v.

.p .r . " r c. n a a s ...v . s.. t A hs u p..g,, l .Lu u i s i / ) .n .w...e.,.c c.c ) ) na u c c y e *. u..o s. a ,., a a r'n' Iw s s.~ 7v. C - CC. slo - ev. m

• a e

--m s rJ L v n.4 3n,.,<- v...r. ;

u. v a.o..c.. - o. r g.

n. h I (Seabrook Station, Units 1 ) and 2) ) ) e- .s.,2 : Cn,,.- O S R,/tes r c s 1 ~ herebs certif, that copies of One " Statement of Dr. M1:hael Channery on Renand to the Atenic Safety and Licensing Appeal Board" were nailed postage pre-paid this 17tn day of Fe'oruary 1951 to tne ~ s

c. c,6. n y *..n v

y* e ? Alan S. Rosenthal, Chairman Dr. John H. Buck A'.e.

1. c

_C a # a_.'f a' 7 4 - a_. s i. - n' e.ic a.#c'e~ & 7 4 -a..s i..~ c m. u m. u -- 3 Appeal Board Appeal Ecard ..a gu. a --.-"; U.S. 'Iu^' e = - "a" a ' = o v Cr--i s s i o n S ^ "v.S. ".. u ' a_ a - ..-3 Co==ission Washington, D.C. 20555 Washington, D.C. 20555 sn- -=- .A e s.' s - = a. '..A ". ~. o -, o_ ;- "a e.. 2 a '.

r. a..k W 4

.k.".. r n o,. g..,... ....c-..a. n,.a a_, v..._: _,, ..a_ ~..c..

o.., o,_ m,._ :

m ;, ; s.;,.. u -...4.,...s._a' o o s_ - ~ 3 '.a. C##'ce c# - "..e n' "_" s- -..a v, G a_ .a_ a ' r, - 4.s.4 - 4 q eo

u.,,a s o n...,. o.s.,

u.,~,.,.. ,n. o u. u-. .s.,-.=; c a.. c-c.: = ;,. a m.:. u.. a_ m..--a, .., _ p a.;.,.m,c1 v .s -.

0.. a_

n' o".".'.~. --.,.. 3.'. e n Boston, Massachu.setts G2103 Thomas G. Dignan, Jr.< Esc.uire Rones & Grav. ..w w w - n. $ a M_...Ma, --gur.PO ~..M.. h.W AF r Q *. *. ~~- u b. 33 w { 8**g [.l.I

• 2, c '...,. a, C, E,, 7 - A.m.

.s -. . - -., - -. =..

• e C^a**=*a"*.**

f 3 S I. [,* a ...t.*_*u*sA. =. c ** h S 1 r -s 4 a 0.* 9 ; S t.~ e e t 3 s.: 3,. r.s0 T.., ;.n ' :.. c. '..a c.. z,. .a..

u. p.. p,.3... p 2

.. <. i.. %;v- ,2 q n, 0 6. w u. :. n, .a_&r.,. : ... y 4 %,. v.. ...., e.lg3. p.n -. 1y.. ct _ -, }, v c.... 4 .a a l v, %.. "6 - w e ') v v e......-m-3,

• '2.w-.4. 0.%-.he-

~.J.e. '.rnt, S=- ) f

• . s 3..

r.g m% J 4e s b j 'M [ '.E y

• I I I

~ " "9 -.e.%# O- ,** g. a.% g. E +. 3 eg,e g o..w s a~_g-. V** - T ~ a O..n g,.,). 3 a c w..y (w, = m. A r= .* o e, 'u..Q. T 4--Ae o ~ as

t.. a ~ b. #. ~y ' "s..,

D.#. 70E_E5. a. m

• Dr.

W. Feed Johnscn D. Pierre G.

Caneron, Jr.,

Esq. ?. t c c. c Safety & Licensing n..;ra;. 2n el e A.O.nc21 30ard w. .c.~.,-- CO ~, 2 3 n 1+ C.' z U.S. 'uclear Regulator' Cor_'.i s s io n -.n.;

-ashire clr 'treet Nashington, D.C.

20555 1000 s ..2 i.,,. ,. r.: e,~e< v-,s

• !s. Ell abeth H.

Weinhold

• Atomic Safety and Licensing 3 Godfrey Avenue Board Panel Hampton, New Hampshire 03842 U.S.

Nuclear Regulatory Commiss: Wasning:an, D.C. 20555 Office of the Attorney General 208 State House Annex Concord, New Hampshire 03301 < c M~ C;;W?5 Wi l l, ' arS. Jordan, III

• h a nd -de l l*/ er e d

=

5

• W M S qu.e 6 em me

@ 4 O' Oh '

• O e

e Ex.M it 1 t / fd "NDCKE7*,3 ('m ';3Nec F[@ ! 7 tga > C t4 ~ as d C4 fica..,g. \\ w g F A Ocese;,,( 4 nM, s, x 4 f kLu ~ / 1 \\/ ~ '%, / iEO' h s l l l l l l

89 EARTHQUAKE NOTES, VOL. XLIV. N05. 3-4, JULY-DECEMSER,1973 Earthquake Statistics in Southern New England Michael A. Chinnery and Donald A. Rogcrs Department of Ccological Sciences, Brown University Providenec, R.I. 02912 ea 0 A93!RACT New England has the longest recorded history of earthquake activity in the United States. Because of the high population den- ,1r sity and because the historical data are likely to be more complete in the Southern New England, we have examined the statistics of the earthquake data and then constructed recurrence relations in an at-tempt to cscinate the mean return period as a function of earth-quake size. INTRCDUCTICN New England has the longest recorded history of earthquake activity in the United States. Several catalogs of earthquakes in this araa have been compiled, the most comprehensive of which appears to be due to Smith (1962, 1966) and covers the period from 1534 to 1959. Smith's data are used throughout this report. No attempt to include information after 1959 has been made, in order to preserve the apparent homogeneity of Smith's data set. It seems very likely that several analyses of these data have been sade in the past. However, if this is so, the results of the studies are not generally available in the scientific literature. Instead, it is common to find, in reports by insurance companies, site investigators, city planners, etc., vague statements concerning the low level of seismicity in } this area, the infrequent occurrence of damaging earthquakes, and even the maximum size of earthquake that may be expected. In view of the high population density in Southern New England, it dees not seem advisable to base major planning decisions on state =ents such as these. Instead, we must examine the historical record in consider-able detail. These data are.far from perfect, but they are essentiallw all that we have. The level of seismicity is low enough that little information can be deduced frem instrumental ;ecords, which are only avail-able after about 1925. In addition, the historical data suggest that the i seismicity observed in the last century. may. be unusually low. The purpose of this paper, then, is to examine Smith's earthquake catalog in detail. We shall concentrate en the Southern New England region, because of the high population density, and because the historical data are likely to be more complete in this area. We shall examine the statistics of the earthquake data, construct recurrence relations, and attempt to estimate the mean return period as a function of earthquake size. We shall e l study both the area as a whole, and also several smaller subareas where much of the historical activity has been concentrated. THE DATA Figure 1 shows Smith's (1966) map of epicenters in the New England area, for the period 1534-1959. This map is a portion of Smith's much larger diagram covering Eastern Canada and the Northeastern United States. In all, Smith lists 729 earthquakes in the Northeastern United States. eNow at California Division of Mines and Geology, Sacramento, Calif. 95814 1

90 EARTHQUAKE NOTE $ fe i tX L, c\\ o W h\\ \\ %de ' _. ) o _l O \\ O o O AN g.\\ O s O M\\ Of \\ 0 '.~ o 'C O a a ) c- . o, t<t n Co 4 O I o a -\\ 'O O

o. D.o..a.

o g o,'. I S/)\\ 'o /,. o o ~ g of, i = T t = s.rv waGuiruct

• k f

..tB o k \\o\\ ]. j s - %@. D \\ d,? *\\ s n\\ E e ~ =3 .\\ - c er (,G ;,* ? g c.e*, o 2_ g, g p-oo o m_-=r OO /

_s

~~ ) .O O sL.., *,_ 8-i In Ae,o o cs a . v g.h vp 3 ty ox }* lf O'

• o O.

= cE ' gd \\ '1_, y st a ., m... m ,0 7 o t -u,, e t m... .... m, g o)rivt EP'et%7ets k[k t C 5C succaaan sta;;c:4 l \\ ua a \\ scatt >= wittA l Fig. 1. Earthquake epicenters in the New England l area,1534-1959 (af ter smith,1966). i We have chosen to select a portion of the New England area fer study. l This portien, which we term " Southern New England," is shown in Fig. 2. Included are the states of Massachusetts, Coemecticut, Rhode Island, and the southern parts of New !!ampshire and Maine. The ocean area East to hasLa p y

EARTHQUAKE NOTES, VOL. XLIV, NOS. 3-4, JULY-DECD15ER,1973 91 0 44 N l B ~ j e I j_ 43 . = >. 7,' A ~ ~.' / d o p- '42 l l i o j j 1~ l / .,+. 1 l C IV.- ha s - r-a i [ tl, 0 41 fd ~ 7b" 71 70"W Fig. 2. The area that we have chosen to define as " Southern ~ New England" for the purposes of this study. ( 69.5* est has been added to these, so that a nu::ber of poorly deterr.ined epicenters off the coast of Massachusetts =ay be included in the statistits. New England (Jefined above) is more distant than 50 miles frem a major cen-ter of population (100,000 inhabitants or ore). In some average sense, therefore, a random epicenter in this area is likely to be within 25 miles of a large population center. This is of sore i=portance in attempting to estimate earthquake risk in the area as a whole. Much of the seismic activity in this region (see Fig.1) is con-t centrated into three zones, which are labelled in Fig. 2 as A, 3, and C. A includes the area around Soston, 3 refers to the southern part of New i Hampshire, and C denotes the togion of Connecticut around Hartford. The I { division between tones A and 3 is rather arbitrary, and the statistics of these zones are analyzed, both separately and together, later in this report. 1

92 EARTHQUAKE NOTES In all, Smith lists 353 epicenters in Southern New England, of these 99 (28*) lie in area A, 96 (27%) lie in area B, and 55 (16%) lie in area C. The three active zones, therefore account for 70% of the data for the whole area. There are three principal errors in this type of information. These are uncertainties in epicenter locations, problems in the determina-tion in intensity, and incompleteness of the data set. The uncertainties in = cst of the epicenter locations shown in Fig. 1 are quite large. The historical data will clearly be strongly influenced by the population distribution. Note, however, that the larger earth-quakes, whose effects extended over a large area, are likely to have less accurate epicenters than small ones. More recent instrurental determina-tions of epicenters are also subject to error, thcugh of a dif ferent kind. The seismic travel time curve in New England is not well known. This is in part due to the heterogeneity of the geology, and in part due to the poor spatial distribution of epicenters in relation to ebservatories in the area. Even the large New Hampshire carthquakes of 1940 (M = 5.S) cannot be located to better than +20 km. It is doubtful whether any of the epicenters on Fig. 1 are any = ore accurate than this, and most are =uch less accurate. The determination of the intensity of an earthquake from historical eyewitness accounts is notoriously dif ficult. Estimates are subject to population distributien, the personal feelings of the observer, and the interpretation of the cataloger. The influence of these factors is mixed. Observers are likely to overestimate the intensity of an earthquake shock. L However, the cataloger has clearly tried to take this into account in his assignment of intensities. In ? 'dition, if the population density vts sparse, it is quite possible that no report was roccived from the highest intensity zone close to the epicenter. In view of this, it does not ap-paar realistic to assume that the historical reports are grossly exagger-ated. In some instances they may result in underestimates of the earth-quake intensity. The worst problem of historical earthquake data is, of course, its completeness. There is no doubt that the data becoees = ore incomplete as one goes further into the past (at a given intensity) and to smaller intensities (at a given eine). On the other. hand, we need the longest time period and the largest range of intensities possible in order to arrive at =eaningful statistics. Because of this, some subjectivity is nesessary in selecting the portions of the data to be analyzed. It seems likely that data regarding earthquakes that occurred before 1700 are unreliable. We have not tried to use these data. How-ever, there is a high probability that all of the large earthquakes since then have been recorded. Similarly, it is enly in the recent past that we may expect a f airly complete record of rnali events. Therefore, in order to try to exclude this type of deficiency from the data, we have chosen to analyte three subsets of each data set. These subsets, shcwing the cime interval studied at each intensity, are listed in Table 1. By com saring the results for the three subsets, we vill, hopefully, obtain seee in!?c=ation about the completeness of the data set as a whole. INTENSITIES AND MAGNITUDES Intensities =entioned in this report refer to the Modified Mercalli Scale of 1931 (see, for example, Smith, 1962). Magnitudes, where quoted. are local magnitudes. Historical eyewitness accounts lead to estimates of the intensity of the earthquake at the observing site. Magnitudes can only be determined reliably from instrumental records. Because of the -t-- w 7 t-F =v e -" '~*

i 4 93 CARTHQUAXE NOTES, VOL. XLIV. NOS. 3-4, JULY-DECEMBER,1973 Table 1. Subdivision of data into subsets. fSubset2 fSubset3 Intensity subset 1 IX 1700-1959 1800-1959 1!60-1959 [ VIII 1700-1959 1800-1959 1860-1959 VII 1700-1959 1800-1959 1360-1959 VI 1300-1959 1800-1959 1860-1959 V 1860-1959 1960-1959 1860-1959 IV 1*00-1959 1900-1959 1900-1959 III 1928-1959 192S-1959 1929-1959 II 1923-1959 192S-1959 1928-1959 nature of the historical data, we shall use intensities throughout. It appears in general to be possible to relate the =ax1=um epi-central intensity I to the local magnitude M by a linear algebraic ex-Gutenberg and Richter (1956) determined the following rela-pressica. tion for Southern California: (1) M=1+jI. The number of earthquakes in the present area for which both M and I I are known is ssall. Figure 3 shows that these data are consistent with i the Gutenberg-Richter relation. A least squares fit to the data points I in Fig. 3 leads to the relation: (2) M = 1.2 + 0.6 I. ( In view of the uncertainties in both I and M the difference between Eqs. 1 and 2 is negligible. The linear relation between I and M is a useful one. Using it we =sy convert instrueental magnitudes into inte.sities (sene recent earth-More in-quakes in Smith's cataleg are listed with only a magnitude). it enables ts to compare our. local statist.cs with global /

portant, studies, which are usually quoted in ter s of magnitude.

It will be convenient at this ata;e to define what we mean by a On the basis of the Modified Mercalli Scale of in-danaging earthquake. tensity, the onset of considerable destructive ability occurs at an epi-We have therefore chosen to l central intensity in the range VIII to IX. define a damaging earthquake as one with an epfcentral intensity of VIII1/2 or greater. This is a very conservative choice, since even an l earthquake of intensity VII is likely to cause some damage, particularly l if it occurs in a heavily populated area. An intensity of' VIII correspends to a nagnitude of about 6.5 l f The possible destructive effects of an earthquake of this size may be Consider-illustrated by the 1971 San Fernando earthquake in California. ( j abic da=4ge and rather high acccleratiens of the ground were observed in the case of this =uch publicized event. In additien to the size of an earthquake, we must also consider the areal extent of the region subject to dasage. This is much less well defined quantity, since it depends so much on the superficial Linehan (1970) has given an earthquake intensity attenuation geology. I I

EARTIK"ME NOTES 91. Vill - / Least Square Fit / / M = l 2 + O 6I 4 Vll -~; / / ./ VI / s. / _m Z /. m v =/ rz f l /

2. iv l

2 Gutenberg - Richter i M = 1 + 2_ I 3 Ill l 2 3 4 5 6 ,M AGNITU D E M Fig. 3. The relationship between magnitude and intensity for 15 earthquakes in Northeastern United States and Eastern Canada. f scale which suggests that intensity VIII will extend out to a radius of about 15 miles frem an epicentral intensity of VIIII/2 This is also a conservative estimate and the radius may easily be doubled in regions of unfavorable geology. In view of this, and the high population density of the ares under consideration, it seems unlikely that an intensity VIIII/2 earthquake could occur in Southern New England without causing considerable dae.sge and loss of lif e. Certainly, any earthquake with intensity greater than this can be relied upon to cause great damage. For this reason, we shall also pay some attention to the possible future occurrence of earthquakes of in-tensity IX and X.

EARTHQUAKE NOTES, VOL. XLIV. NCS. 3-4, /ULY-DECDf3ER,1973 95 FREQUENCY-INTENSITY RELATICNS It has been clearly demcastrated in many parta of the world that there is a linear relationship between earthquake frequency and earth-quake =agnitude (see, for example, Evernden,1970) of the folicwing form: log N - a - bM, (3) g where N is the number of earthquakes occurring within a region in a given time period with a magnitude greater than or equal to M. a and b are con-stants; a depends on the size of the ares chosen and the length of the time period concerned, and is an overall measure of the seismicity of the area. b usually lies in the range 0.5 - 2.0, and appears to be related to the nature of the tectonic activity causing the earthquakes. Logarithms, un-less et rvise stated, are to base 10. If a linear relation exists between magnitude and intensity, as we have discussed earlier, then clearly se may write (4) log N, = c - dl, where, now, N is the number of earthquakes occurring within a region in a e given time interval with an intensity greater than or equal to I. e and d are constants. The Relation 4 is a very useful ene. It enables us to use the data for smaller earthquakes, which are plentiful, to determine the frequency of occurrence of large earthquakes. In the later sections of this report we shall attempt to determine the constants e ar.d d from the data in Smith's catalog. l In considering the statistics of the earthquakes, instead of using the quantity N:, it is scre convenient to define the "zcan recurrence time" (MRT). MRT is simply the aversge time between earthquakes with a given intensity I or greater, and is equal to 1/N time periods. Ve shall e be particularly concerned with the deter =ination of MRT for dasaging earth-quakes. Before we proceed, however, we must 'honsider the range of validity of Eq. 4 Where complete data has been obtained, the frequency-magnitude relation (Eq. 3) has been shown to be valid over a remarkable range of =agnitude (from greater than 8 down to less than 0). There is some theo-re Acal reason to suspect that there is a limit to the possible size of earthquakes. If this limit exists. it is not well known, and may be of the. i order of magnitude 9. Such theoretical linics are well beyond the sizes of the earthquakes that we shall consider in tids report. We must next examina whether there is any evidence that there is en upper limit to the sizes of earthquakes to be expe' ted in the New England e l' area. There seems to be some confusten on this point. In fact, there is no basis for suggesting that ruch an upper limit exists, and as we shall see, analysis of the historical data supperts this statement..The largest earthquskes that have been recorded in Southern New England are listed in l lesst one, and probably two earthquakes of intensity IX (mag-l Table 2. At nitude about 7) have been recorded in the past 400 years. This length of f record is far too short to conclude that an event with intensity X (or i greater) has not occurred in the past.,or will not occur in the future. The chseleston, South Carolina, earthquake of 1886 had an intensity X, and l occurred in an area that is somewhat less seismically active than New England. l l l l

96 EARIUQUAXE NCTIE3 Table 2. I.arge earthquakes in Southern New England. l Locatica Intensity Date 1568 Rhode Island VII 1574 Rhode Island VII 1584 Rhode Island VII 1592 Phode Island VII July 11, 1638 Off Cape Ann. Mass. VIII November 9, 1727 Near Newbury, Mass. IX June 14, 1744 off Cape Ann, N ss. VIII November 18, 1755 "about 200 miles" East IX of Cape Ann, Mss. May 16 or 18, 1791 Near Moodus, Conn. VIII October 5, 1817 Northeastern Mass. VII December 20, 1940 Ossipee Lake, N.H. V!1 December 24, 1940 Ossipee I.ake, N.H. VII ~" 'Je mus t therefors admit the probability that large earthquakes will occur in Southern New England, if at infraquent intervals, until some new infor:ation arises that dis =isses this possibility. It should be added that the absence of a very large earthquake in the recorded history of Southern New England is not a reason for cocplacency. It is conceivaole that a long tir4 has el:psed since the last large earthquake in this area. If this were the case, the probability of one occurring in the near future could be quite high. REC"PJtENCE RE1ATIONS: SOUTP.ERN NEW ENGIXiD 'Je consider first the whole Southern New England region (defined in Fig. 3). Smith (1962,1966) lists 353 events in this area during the period 153'-1959, after all obvious aftershocis are removed from the data. The distribution of these earthquakes in intensity and cine is shown in Table 3. i.*here the intensity of an event is listed as being between two levels (e.g., IV-v), one half event has been included into each level. Table 3. Earthquake data for Southern New England. I Intensity Before 1700-1520-1860-1900-1928-1700 1799 1359 1899 1927 1959 2 IX VIII 1 2 2 VII 4 1 = VI 1 2 3 1/2 1 1-1/2 V 2 8 5 6-1/2 9 6-1/2 IV 5 16 24 13 22 21 III 2 16 40 23 14 26-1/2 II 3 3 25 5 32-1/2 It is clear from Table 3 that the data from before 1700 are very incoeplete. At the lever intensity Icvels this incompleteness con-tinues until late in the historical record. For this reason, we have disregarded portions of the data, and have analyzed the remainder in 1 9'

i I i l I 97 EARTHQL' AXE NOTES, VOL. XLIV, NCS. 3-4, J"f,Y-DECDtBER,1973 i Southern 25 New England 20 Log Ne ~- Log Ne :.t 30 - 0 57 I 15 j r yo 1 i i 05 o 0 - il IV VI Vill X INTENSITY I fig. 4. Frequency.-intensity plot for 135 events in Southern is the cumulative nusbar of events New Fngland. Ne (with intensity I or greater) per century. Note, however, that all the three subsets described earlier (Tab 7sr 1). These large earthquakes with intensity VIII or IX occurred before ISOC. events will therefore only appear in subset 1 of the data. Frequency-intensity plots for the three subsets of the acta have As may be expected, the large events in the 17GC's been constructed. e conclude that subset 1 is unreliable. make subset 1 very nonlinear. identical, and we therefore have chosen to use Subsets : and 3 are almost The subset 2 (which contains more events) as our most reliable data set. The ordi-frequency-intensity graph for this subset is shown in Fig. 4 nate is the logarithm (to base 10) of the cumulative number of events with intensity 1 or greater, per century. The low points The points in Fig. 4 define a fairly linear relation. at intensities II and III are to be expected, since it is virtually 1:n-I possible to obtain a complete record of these small events, even in theT recent past. For this reason, and because of the results within the range of 0.54-0.60. given in the next section of this paper, our best estirate of the slope of The data then determine the frequency-intensity relation is 0.57 ($.03).

T l 98 EARTHQUAKE NOTES the following recurrence relation: Log N, = 4.30 (+0.15) - 0.57 (:0.03) I, (5) converting this into a frequency-magnitude relation; using Eq. 2, we obtain: Log N, = 5.45 (+0.20) - 0.95 (+0.05) M. (6) The "b-value" in the range 0.9 - 1.0 is very reasonable for an area ,e such as New England. b values lying in the range 0.8 - 1.0 are found in most parts of the world (Everndren,1970). Isacks and Oliver (1964) found a b value of 0.9 in their study of small earthquakes recorded instrumental-ly in New Jersey. The errors quoted in Eqs. 5 and 6 are based only on the fit of a linear relationship to the data points. They do not inclade contributions frca errors in the data points themselves, which are extremely hard to estimate. RICL*RRENCE REIATIONS: 3)STCN-NEW MAMPSHIRE RECION Areas A and 3 combined (see Tig. 2) irclude the Boston vicinity, Northeastern Massachusetts and ca. ?ssocia<2d offshore region, and the Southern half of New Hampshire. Smith lists 194 events in this active zone, which therefore accounts for just about 50 of the total activity in Sauttern New England. The distribution of these events in time and intensity is shown in Table 4 Table 4 Earthquake data for Boston-Southern New Hampshire regicn (areas A and 3 cembined). f1860-Intensity Before 1700-1800-1900-1928-1700 1799 1859 { 1899 1927 1959 IX 2 VIt1 1 1 VII 1 2 VI 1 t 1/2 -1 V 2 6 -2 4 7 1 l I" 4 6 '12 7-1/2 9 8-1s? ' III 2 16 16 12 6 13-1/2 II 3 3 21 3 16 i I l Frequency-intensity plots for these various subsets of these data show very similar features to those found in the previous section for the l whole of Southern New England. Inclusion of the ez 'y data (subset 1) l r leads to a very nonlinear plot. Subset 3 shows ssme scatter due to an l{ insufficient number of events. Subset again gives the most reliable l1 data set, and the resulting plot is shown in Fig. 5. Linear relationships fitted to the data again have slopes 'a the range 0.34 to 0.60. This is an important point, f or two reascn s. t Firstly, it substantiates the sirpe deter =ined for the Southern New England region as a whole. Secondly, and more importantly, it strongly suggests that the slope (or b-value) is roughly constant throughout the area under study, within the resolutien of the present data. As before, then, we assume a slope of 0.57 (+0.01) for the fre- } quency-intensity plot. This leads to the following racurrence relation: I l l [ l l l u.

99 EART!!QUAXE NOTES, VOL. XLIV. NOS. 3-4, JULY-DECEP3ER,1973 25 Boston - New Hampshire Region 2-0 log ~~ N e 15 Log Ne: 4 00 - 057I 10 0-5 0 ? f 1 T f 1 e t g 11 IV VI Vill X INTENSITY I Fig. 5. Frequency-intensity plot for 65 events in areas A and 3 (see Fig. 2), which include the Boston vicinity and Southern New Hampshire. N is the cumulative number of e events (with intensity I or greater) per century, e Log N = 4.00 (+0.15) - 0.5 7 (+0.03,P I. (7) c And, using Eq. 2, ve find Log N, = 5.15 (+0.20) - 0.95 (_+0.05) M. (3) RECURRENCY RELATIONS: AREAS A, B, C Subdivision of the Bosten-New Hampshire region into the individual subarcas A and 3 starts to point cut scee of the inadequacies of the his-I torical data set. Surerficially, 5:nith's catalogue includes 9d-1/2 events in ares A, and 95-1/2 events in area S. One is te pred to ascribe one-half of the activity in the Boston-New Hampshire region to area A, and cric-half to ares B. Ilovever, tabulation of the events in these two areas as functions of time and intensity shows up some marked differences. Table 5 shows this

P 9 r. 100 EARTHQi'AKE NOTf5 Table 5. Earthquake data for Besten vicinity (area A). Before 17 W 18]O-1360-1900-1923-ntensity 1700 1799 1859 18 0 1927 1959 IX 2 VIII 1 1 VII 1 VI 1 2 2 V 2 5 1 1-1/2 5 1/2 IV 4 5 9 1 5 3-1/2 III 2 13 4 3 5 1-1/2 II 3 3 5 3 3 Table 6. Earthquake data for Southern New Hampshire (area 3). 5efere 1700-1800-1960-1900-1928-Intensity 1700 1799 1839 1999 1927 1959 IX VIII VII 2 VI 1/2 1 V 1 1 2-1/2 2 1/2 IV 1 3 6 4 5 III 3 12 9 1 12 II 16 13 tabulatica for area A, and Table 6 shows the same for ares B. Area A ap-pears to have had a relatively high activity in the 1700's, which has since been steadily decreasing. On the other hand, area 3 appears to show a icw in activity in the 1700's, which has been increasing since then. The reality of this difference is, of course, questionable. It is likely that the New Hampshire data havd' been heavily influenced by the effects of population distribution, and that the earlier parts of this data set are very incoeplete. This raises an interesting question. It is clear that combining the data from areas A and B leads to an estimate for the seismic activity that has been relatively cnifors since the 1700's t (see Table 4). Is this apparent uniformity real? It is worth while zen-tiening the folicwing possibilities: f*.) The area A c 2ta may be fairly reliable, while area 3 may be very incomplete. Addition of the "=issing" New Hampsnire events will bias all l our recurrence relaticas in the direction of increased seismic activity. os lf If this is the case, we have a strong indication that the seismic activity during the past 100 years or so has been anusually low. This may be the j result of the statistical fluctuation, or some unknown physical process. (ii) The early high activity in area A =ay be the result of exagger-sted intensity estimatas for some of the events in the 1700's. If this is j so, it is possible that the activity cf the two areas has been relatively j unifors throughout the historical period. i l

1 I EARTHQUAXE NOTES, VOL. XLIV, NOS. 3-4, JULY-DECC'3ER,1973 101 i j There is no way to distieguish between these possibilities using the present Jate. I!owever, the second possibility will clearly lead to the most conservative estientes for the seismic activity. Direct construction i. of frequcocy-intensity plots leadJ to incanclusive results because of the small number of uscable cysnts. be therefore return to our first inclina-f tion, ar'd.tsaume that the scismic sctivity of the Boston-New Hampshire region las evenly divided between areas A and 3. This leads to the follov. l ing estimates for the recurrence relations in areas A and 3: 32

j' Log N, = 3.70 (+0.15) - 0.57 (p.03) I, (9) o Log N, = 4.85 (f0.20) - 0.95 (3 05) M.

(10) i Clear;y, the quoted errors are not a true reflection of the possible in. accuracies in these relations, which may be considerable. They may, how-ever, give a sore reliable esti ate for the lower limit of seismic activity in the two areas. Relatively few events have been recorded in the Hartford, Connecti-cut, vicinity, denoted as trea C. Smith lists a total of 55 events in this area, of which only 20 fall in subset 1. Utis number is quite in-adequate for any statistical treatment. '.e say, however, get a rough idea of the activity in this region by assuning that the slope of the frequency-intensity is knew (0.571 0.03), and that the record of events with intensity IV is ecaplete during f the period 1900-1959. l This is sufficient to deter =ine the following recurrence relaticas f for area C: Log N = 3.35 (3 20) - 0.57 (10.03) I. (11) Log N, = 4.50 ($.25) - 0.95 (10.05) M. (12) MEAN EECURRENCE TIMIS yrom the recurrence reistions listed in Eqs. 5 throuah 12, it is easy to eniculste the mean recurrence tires. These are listed for e g i ; variety of intensities in Table 7. It should be noted that these were i determined frca the cumulative event frequencies. Thus the first entry in Table 7 states thJC thC 2eam interval betveen earthquakes with in- [ tensity VIII g creater, in Southern New England is about 180 years. f Table 7. Mean recurrence times'(in years) t i i in u, l u.m. fi w - u..,. I..... l ........... I... c. I t ..e %.m etts

4. w-4. 3 lae Q.8)

> ';w) '90 Q2000 .oe QMel 1600 Q.GBl st ri.an

o. )-4. 7 we (.aes m q:m nao q.,si isos q.ooi 3ces gionen e

1: .. r. o we g:nce t.w q.ooi .s.= qwoot two owoo: 600e graset 1

7. 2+ F. I DAB Q600p>

SAS Q:Not woue Q.ft9) ISoce ( coc) 32000 (.e0003 e It is interesting to coepare these =ean recurrence times with the ti=es since the last large earthquakes in the area (Table 2). The last l earthquake listed with incensity VIII occurred in 1791, just ISO years i ago. Clearly, reprdless of the mothed used to epiculate future proba-i! bilities. another earthquake of this size may be expected in.che near a I i - [.

102 EAATHQUAKE NOTES future. The subject of the determination of earthquake risk f rom these data vill be taken us in a later paper. It is worth erphasizing that the mean recurrence times and recut-rence relations were calculic 4d without using the large events (I > VIII) in Taole 2. They aro therefore independent of any errors in the intensity estinates for these large events. CONCLUSION 3 ~ The principal conclusions of this study say be summarized as folic +s: I 1. The data in the Smith catalogs are consistent with a "b-value" of 0.95 (3 05), applicable both to the Southern New England area 0 as a shole, and also to smaller regions within this area. 2. Recurrence relations for the whole area and for certain sub-areas are listed in Eqs. 5 through 12. 3. Southern New England is likely to experience an earthquake with intensity VIII or greater in the f airly near future. The mean re-currence time for events of this size is about 150 years, shile the last event of this size occurred just ISO years ago. 4 Of the total activity of Southern New England, approximately one half is concentrated in the Bosten-Southern New Hampshire region (areas A and B in Fig. 2). The remainder is scatrered throughout Southern New England, with a minor concentration in central Connecticut. 5. There is no evidence to suggest that there is any upper limit to the size of the earthquakes that may be expected within this area. Earthquakes of the savority of the Charleston, South Carolina, eartt. quake of 1886 (sagnitude about 7.5, intensity about x) probably occur in Southern New England with a mean recurrence time of several thousand of i There is no historical evidence to suggest when the last event years. of this si:e cccurred. 6. Most of the large earthquakes in this area occurred during the It is not clear if that century was unusually active, or Isch century. if the last 200 years has been unusually quiet. All of the statistical conclusions in this report have been based on the data after the year 1300, and therefore do not include this earlier high activity. It is therefore possible that we have underestimated the activity of the i l area. i i REFERENCES Evernden J. F., 1970, Study of regional seismicity and associated l problens, Bull. Seism. Soc. Am., 60, 393-446. l f Gutenberg, B., and Richter, C. F.,1956 Earthquake magnitude, in-tensity, energy, and acceleration. Sull. Seism. Soc. Am., 36,, 105-145. Isacks. B., and Oliver, J., 1964, Seismic waves with frequencies from 1 to 100 cycles oer second recorded in a deep mine in Northern New Jersey, Bull. Setsm. Soc._ g, 3,1941-1979.

Linehan, D., S.

J., 1970, Geological and seismological factors in-fluencing the assessment-of a seismic threat to nuclear reactors, in seismic resi. n_for Nuclear Power Plants. R. J. IIansen Ed., The M.I.T. Prese. 67-90. "^ m.

EAAT!!QL'AKg,907g5, yng, g gy NOS 3-4, L'LY-DECEM3ER, 1973 103 e I8 and adjacent areas, $$a.**6,.71301. $em 53-1957 b Pu - ~ Smith, 'd. E. T.,1756, E4 rthquakes of eastern Canada aad adjacen: areas, 1928-1959, Publ. Dom. Obs. Ottawa, 3,67-121. 5 O 4 e 4 O e. m i e i 0 i l

6 I l r I I l. I, ~ 0, Exhibit 2 1 i 9 1 I h l

b I Bui.ieun of the Seismoi.mcal S rwty of A enca. Von 69. No. 3. pp 75* 4*2. Jwr'e 19""J A COh1PARISON OF THE SEISSIICITY OF THREE REGIONS OF THE EASTERN U.S.* BY }{lCHAEL A. CHINNERY ABsTR4CT Frequency. intensity data from the Southeastern U.S., Central %ssassippi Valley, and Southern New England are compared. They are all quite parallet to or.e anoteer and consistent with a slope of about 0.57. There is no evidence for the existence of upper bounds to maximum epicentral intens6ty in these data sets. Linear extrapolation of the frequency-4ntensity data tointensities of X leads to expected probabilities for the occurrence of large earthquakes. The largest everts which have occurred in these three regions are cor44 stent vth these procabilities. INTacoccTtoN Recently there have been rather detailed analyses of the seismicity of three sections of the Central and Eastern U.S. Bollinger (19731 has desenbed an extensive set of data for the Southeastern U.S., which includes the seismically active zenes of N'aryland, Virginia, West Virgima, North and South Caro' na, Georgia, Alabama, d and Tennessee, for the period 1754 to 1970. Nuttli (1974) has listed the known events in the central 3fississippi Valley seismic region for the period 1833 to 1972. And Chinnery and Rodgers (1973) have analyzed the data of Smith (1962,19661 for the Southern New England region for the period 1534 to 1959. The purpose of this ps per l is to compare these three studies, and to bring out the similarities between them. The discussion of seismic risk inevitably involves plotting frequency-intensity (i.e.,

j manmum epicentra' intensity) diagrams. In what follows we use this type of plot, j

since magnitude data are not available for all three regions. This raises a difficult point, since within each of these regions, the seismic activity is not uniform. The 2 f selection of the boundaries of the area to be studied is much akin to the problem of i the definition of a tectonic province (which is required, for example, by the Nuclear j Regulatory Commission Rules and Regulations, Part 100,* Appendix A). For the moment, we shall nake the following assumptions: First, we assume that all subregions within a given region have a linear frequency. intensity relation of the h' form l l log N, = a - bl l l l where N, is the cumulative number of events in the ith subregion with intensities j greater than or equal to I, and a, is a parameter desenbing the level of seismic

l activity of the ith subregion. We assume that the slope b is common to all subregions, Second, we assume that the maximum possible intensity in each subregion,if one i',

exists which is lower than the nominal maximum of XII, is larger than the largest j event recorded within that subregion during the period of the earthquake record. These assumptions sound very drastic, yet they are really implicit whenever we plot a frequency-magnitude or frequency-intensity curve. Furthermore, at least in l l l

• The v'.ews and conclusions contained in this document are those of the contractor and should not be interpreted as necessardy representing the official pohetes. either expressed or tmphed, of the United States Govemment.

757 l

758 MICHAEL. A. CHINNERY principle, they are testable. It is easy to plot frequency-intensity diagrams for portions of a region and examine both the linearity of the results and the constancy of the slope b *n practice, of course, scatter in the data often makes such a test inconclusive. However, a substantial breakdown of any of the above assumptions should be apparent in the data for the region as a whole, either by the appearance of nonlinearity in the frequency. intensity statistics, or by variations in estimates of b using different data sets. As we examine and compare the seismicity of the three areas under consideration, we shall look for information related to these assump-tions. Perhaps tha most important question which we shall address is as follows Each hese rn.a nr 5ad one moderately large earthquake in its recorded history (the c 1755 Cape Anne,. A11-1812 New Madnd, and 1886 Charleston events). Are these large events consiste..t with the record of emaller earthquakes that have occurred more recenti,7 "...aly, this question has a direct bearing on the very fundamental problem of how to extrapolate from a short record of seismicity to the occurrence of low probability events, which is particularly important in the assessment of the potential seismic trazard to crincal structures such as nuclear power plants. We shall disregard questions of the lack of stationarity of the earthquake process in these three areas, in spite of their potential imnortrace IShakal and Toksoz, 1977). It is very difGcult to document this nonstationarity within tima periods of 100 to 150 years, because of the small number of events concemel THE DATA Southeartern U.S. Bellinger (1973) describes the seismicity of four seismic rones in the Southeastern U.S. for the period 1754 to 1970 (see Figure 1). In this study we shall restrict ourselves to the two southernmost zones, the Southern Appalachian I seismic zone and the South Carobna-Georgia seismic zone. The combined area of ( these two zones is given by Bollinger to be 307,000 km Since we wot,ld like to 2 l exclude the 1886 Charleston earthquake from consideration, we have analyzed l events du-ing the period 190G to 1969. Even this period is probably too long for the adequate reording of intensity 111 events, so these have been accumulated for the period 1930 to 1969 only. Total events listed by Bollinger (1973) are shown in Table 1. These data are easily converted into a cumulative frequency intensity plot, and this is shown in Figure 2. The usual interpretation of such a diagram is to fit the l data points with a straight line, recognmng that the data at the lower intensities is likely to be incomplete. Such a fit is shown as the solid line in Figure 2. This line corresponds to the equation log N, - 2.31 - 0.46I. (1) The slope of this line is low compared to other simdar regions, as we shall see *,elow. l The occurrence of three intensity VIII events during this 70 year period seems high, l and in fact one of them has been shown to be an explosion (G. A. Bollinger, personal l communication). Certainly a line such as the dashed line in Figure 2, which has the equation l log N, = 2.88 - 0.557 (2) i cannot be ruled out. The slope of 0.55 in this equation is very close to the slope 0.56 i l i l .... -..~...... - l l l

.. ~ a e= t 1 l ' NO. VA. -' M D', ~'* I~~ ~ ~ SiEIS WIC : ZONE A a, 'Q l '.-.-qM;f e / s = o, g s , 7. l ..W Vo

-)

)Q r ,:o = i f

o 1

5 S0.dPPALCHIAN..... D* 'l 6 ' 7 i SEISMIC Z,ONE,"~'^ ) i O -4 o a j ~ ' O,. i O L M O o p

• .O

] { . o q Ih -( N C '" g l O SS l13 7 M .. ) q' q LEGEND 1 a-e l W.. 9 e'..g.og, e,- y r.ii siepoei, o g a i 4-lM . e n - in o e o

• ' +

w. w ,y g .4 .y; vi vn 9 => se o Mess ,g..

y

. 4, viti-in m %d, _.. g 2 x Tag av moet 4 M i 'SO CAR LINA-wm* ? f i7 GEORGIA N 4 SEISMIC ZONE go * '" 7" 5 l \\; 7_. o ~ M l. ~ c O f40. I. Selaitthily in the Soullaraaleift II."I Inf Rite l*ti04l I770 la IS$ l(cpsualured, WHh prtInision, huan llulhngtg ( $378 -W-

r 760 MICHAEt. A. CHINNERY

0 06 found by Bollinger (1973> for the whole Southeastern U.S. For the moment, we will retain both equations (1) and (2) as possible interpretations of the data.

Central.Utssissippi Valley. Nuttli (1974) has given a list of events in the central Mississippi Valley for the period 1833 to 1972. The epicenters of these events are shown in Figure 3. The total area of this zone is given by Nuttli to be 250.000 km. 2 Since he lists few events before 1840, w e have restricted ourselves to the period 1840 to 19ti9. Table 2 lists the events during this period as a function of intensity. As TABLE 1 Evt>Ts IN SocTHans AFPA1.ACHI AM OD SOUTH Canous 4-G ronc 4 Setswie ZowEs w-.o P. w M..s a.- 1930-1969 10 IV 19.O-!%9 49 V 1941%9 44 VI 1941%9 17 VII 19Al%9 3 V!II

9n-L%s 3

CS SOUTH CARCUNA-GECPG.A ANO \\ sco Amu.cun \\ SE15Mic DES 0 "" 19C0 - 1969 l \\ l x j 0 -o 5,- e i x 5 z" -i o - l \\ s I \\ I \\ -i s-Log N = 2 31 - O d6I( I g A -2 c - ' Log N. 2.88 - 0.551 t 8 8 t i! I l WTENSITY Fic. 2. Cumulauve hequency.mtensity plot for the data m Table 1. Two pnanible straight lme mterpretations are shown. l before. Smaller events are only counted for the more recent portion of this time penod. Since many events are listed with intensities intermediate between two values (such as III to IV), where this occurs one-half event has been accumulated into each value. This accounts for the fractional events listed in Table 2. Figure 4 shows a cumulative frequency-intensity plot for the data in Table 2. A reasonable linearity is obtained, corresponding to the equation log N, = 2.77 - 0.55I. (3) i

y l 1 SEISMICITY COMPARISON-THREE REGIONS OF THE PASTERN U.S. 761 Southern New England. The seismicity of Southern New England has been discusud by Chinnery and Rodgers (1973), using data of Smith (1962,1966 for the period 1534 to 1959. The region defined as Southern New England is shown by the solid line in Figure 5, which abo shows the epicenters in Smith's listing. Following Chinnery and Rodgers (1973), we note that many of the listed epicenters are s c. ,.s 9 l l

• o. =i=* e x s o....s s
• '0**.

..--s...............3.................... 3r /;. =f.,;i . 1 > v ..,, / c. 7 6:.e i

u 4

.c m... j....,. g. -...... L.....z..b......... /h'.,7 7 I / ..I N% '/ N}. i I a 's / ** j' y'. lt l / ,/ 3,.

5, i

f 4Y r' */O g r.,W,,s%......l......".'..; ,,.f -.. -..a.,2 ..e ll ..). ':... ).# .. e,, ,,i /,- -,)..\\ ,1 l s.~.....,.. . 7 .m. i .rp. .w.w d l 1 5 3.; i / l t.R K l ay,=. / *r lTENN '/ t. OL.- C* C. ,,e i / uo 4* m.....j....A. '"h. .,/ : wss,i I, i ,26-y..\\.'-.* i g6* Fic. 3. Epicenters in the central Mimssippi Valley region, for the pened 1803 to 1972. Reproduced, with permiss!on, from Nuttii (1974). TABLE 2 EvrNTs is CrNTut Mtsarmeri Vutry interumsy P.ense NaadF..ata 1930 198 22.5 III 1986 1% 9 94 5 IV 1370 1% 9 143.5 V 1870 1 % 9 63 0 V' 14M%9 31.5 19).1%9 10 5 'i U: 140.I'h9 10 1NO.i%9 10 clustered in a region extending from Boston through central New Hampshire. We have outlined this area in Figure 5, and refer to it as the Boston.New Hampshire 2 seismic zone. The areas of the two zones in Figure 5 are approximately 100,000 km iSouthern New England) and 27.000 km' (Boston.New Hampshire zone). Since we wuh to exclude the 1755 Cape Anne earthquake from the data set, events have been I

-..,..,.._,....y .,a. f..,.. 762 .utcHAEt, o CHINNERY accumulated in both the Southern New England region and the Boston New Hampshire zone for the penod 1500 to 1959. These are listed in Tables 3 and 4, respectively. As before, small events are only accumulated for the most recent portion of the record. The cumulative frequency intensity plot for Southern New England is shown in Figure 6. The straight line through the data has the form Log N, - 2.36 - 0.59I. (4) In spite of the rather low numbers of events, this line is a reascnable fit to the data. In the case of the Boston New Hampshire zone. however, the number of events io. RSS$$1 ppt Vt.U.,EY 1840 -1969 osL I i low i W z -os - U z* Q 7 -to!w a ~ LoQ N = 2.77-0.55 I C -tr - I 2 22 22 1 NTL'NSITY Fic. 4. Cumulative frequency.mtensity plot for the data in Table 2. becomes low enough that it becomes dif6 cult to formulate a linear fit with any certainty. A straight line through the upper four data points has a shallow slope (about 0.50s, which is signincantly different from the other areas studied, and which leacs to high estimates of nsk for !crge events. We prefer to interpret these data with a line such as the one shown, which has the equation log N, = 2.15 - 0.59I. (5) With this interpretation, the number of intensity VII earthquakes is anomalously high, due either to poor data or a statistical fluctuation. At least equation (5) should lead to reasonably conservative estimates for risk at high intensity levels. i k j

^ ~~ ' ' ' ~ ' ~ ~ O ~ SE!SMICITY COMPARISON-THREE REGIONS OF THE E, ASTERN U.S. 763 [ f // y* \\ ) \\ %.m o o ,4 -)7 \\ w k, O O' l \\ C'\\ k @' \\ qa ~ \\ r\\ Q ea\\ O O h e a O O cm ' o ,\\ ?y 1 s -,-u O ~ N&'% te / me. et a \\ SNTEN5 TY v1GNITUDE 7, j 4*y,g g p t;..g 9 m 1 t

1. g, y

.g a. \\ \\ 1 -m.

• \\

s en O %Lo2.- \\

1. ,a'

n3

\\ g,, = Tu - ,t \\ l% l \\ w f ) l \\a* e O 'h .\\ .. _... 6 o 3 =h-e o g [1 = -h

• c..L., s

\\\\ ir v i\\ C O\\ S' \\a To \\ x-Wo 3 =T o\\of oY \\ O 3 '. c ce M' I f ^ .h ,s / l -t st-i e EPTE %T AE (vacenset.est 1201 o y O E mcE Nin E gy.cene.ne es t t 201 6 olovE EpicE%f mE5 j \\ -O \\ Q SEiSMcGas p t'ariou i o So 100 \\ scatt im witEs \\ l Fro 5. Epicenters m New England. from Smith ilo%. The solid line outhnes the region called Southern New Erziand m thLs study. The broken line indicates the Boston-New Harnpshire zone (see l Chinnery and Rodgers.1973L COMPARISON OF FREQUENCY-INTEN5ITY DATA l The frequency-intensity data shown in Figures 2,4,6, and 7 are shown together l in Figure 6. In this case we have omitted the individual interpretation using fitted straight lines, and show the data alone. This emphasizes the very similar character of the four recurrence curves. There is some scatter, but each of the curves is

764 MICHAEL A. CHINNERY consistent with a slope somewhere in the range 0.55 to 0.60, and we show a slope of 0.57 which seems to be a reasonable average. In view of the rather inferior quality of much historical intensity data, it is surprising how consistent the slopes of cumulative frequency-intensity date appear TABLEJ Evt.sts :N SouturnN New ENcLAND Imeww Perwas Neo d >.*ents II 19 % 1953 32.5 111 1923-1959 26.5 [V 1900-1959 43.0 V 1660-1959 24 0 VI 1800-1959 6/) V11 1&J0-1959 3.0 TABLE 4 EventsN Bostow New Haursmac Zows imeo. rw w a >:.eer, 1928-1959 16.0 III la23-1959 13.5 IV 1900-1569 17.5 V 1560-1 % 9 12.0 VI !!ic-1959 3.5 Vil 1800-1959 3.o ir 'A SON NOW LTJNO 18co - 1959 03 - e e

• ~

t.oq N = 2.36 -o 591 g Or [ -o s ,v 9 I a -so- - t s'- -t o - l t t t t 1 2 a n 1 NTENstTY Fic. 6. Cumulauve frequency-intensity plot for the data in Table 3. to be. Both Connell and Merz (1975) and Veneziano (1975) have surveyed a rumber of estimates of this slope, and many of these are consistent with the present data. The mean of the 11 estimates quoted by Veneziano is 0.53, *.ut his lin contains some low values which are probably not realistic. Of particular in erest are the

F SEISMIC 1(Y COMPARISON-THREE REGIONS OF THE EASTERN t.*.3. 765 values 0.59 for the whole U.S. (Connell and Merz,1975) and 0.M for Cahfornia ( Algermusen,1969). A recent estunate for the area around the Ramapo fault in New York and New Jersey is 0.55 = 0.02 (Aggarwal and Sy kes.197el. It is interesting to compare a slope of 0.57 with the value that one would predict from known magmtude-intensity relationships. A selection of these relationships have been given by Veneziano (1975),in the form 3f = ai + a:I. (6) Values of the constant as have been estimated as 0.67 (Gutenberg and Richter, 1956), 0.69 (Algermissen,1969), and 0.60 (Chinnery and Rodgers,1973. Howell, b 0S BCSICN-NEW HAMPSHIPE 1800 -1959

• ?

e l Loq N = 2.15 - 0 591 e C - 0 5 '- [ s = l -10l-w1 .J e e 8 } t t t I INTENS*TY Fic. 7. Cumulauve frequency. intensity plot for the data m Table 4 1973). The latter estimates of 0.60 were obtained from data in the Eastern U.S., and may be the best estimates for our present purposes. There is an abdunance of frequency-magrQude data, which is usually represented by the form log N, - a - b3f (7) where the slope b often lies between 0.9 and 1.0 tsee, for example Chinnery and North,1975). Combtning this expression with equation 16i. with a, = 0 60, would lead to a slop-of the frecuency-intensity relation between 0.54 and 0.60. Clearly the

e - m smm I-766 MICHAEL A. CHINNERY o u,55.S$,88* <auff a W TMEASTESN'.5 e Wr=ERN NE* E%UhO \\4 OS~ 90STON - Nte =Auss,.et I f Os \\ l 3 \\ z \\ i \\ l r \\ -os-saat o st \\ z h-g 7 t . tow I \\\\ \\ \\ t 1;- \\ l l 23 X hTE% $ITY Frc. 3. Comparuon of ths frequency intensity data from Figu.res 2,4, and 7. l t l o WS$lSS. Pet vau.fr i, L.

• SCUTHEASTE*N v S-e SCUTP*E8N NE4 ENG u NO

& SCSTCN-NE* *PSMPE \\ 20m i

• l

= t. E 25-s l k i l y 30 - i g u l ,9 6a 33 - g E i nl I k l eC-i l 4s-I 1 1 NTENSTv l i Fic. 9. The same data uwd in F:gure 9. but normalized for the area of the vanems zones. I

SElsMic!TY COMPARISON-THREE REGIONS OF THE EASTEMN t'.S. 767 0.57 value shown in Figure 8 is eminently reasonable and con'sistent with other information. The similanty between the four sets of data shown in Figure 8 can be further emphasized by normalizing for the areas of the seismic : gions. After this normali-zation. Figure 9, the recurrence curves are found to lie almost on top of one another (we have chosen to normalize to 1,006 km, but this choice is completely arbitrary) 2 The apparent similarity in seismic activity per unit area is entirely fortuitous, and. is simply due to the particular regions chosen for each study. The tme levels of activity in the three regions differ markedly (see, for example, the return periods calculated in Table 5). However, one is tempted to note that the activity per unit area in the Boston-New Hampshire zone is slightly larger than that in the South-eastern U.S. Is there really any good recon why an event the size of the Charleston earthquake could not occur in the Ba r.on-New Hampshire zone? It is interesting to search thtse data wts for evidence that there may be an upper bound intensity in some of these areas. Cornell and Merz (1975), for example, have prcposed a frequency-intensity curve for a site in the Bor.on area that curves downward and becomes vertical ; parallel to the ordinate axis) close to intensity VII. Since this calculation is for a single site, it is crucially dependent on our ability to predict the location of large events near Boston. Certainly, if large events could occur anyu here within the Boston New Hampshire zone, the present data show no mdications of an upper bound. Given our present knowledge concerning the mech-anisms oflarge events in regions like the Boston New Hampshire zone, it does not seem reasonable to propose such an upper bound. RANDOMNESS OF.ME CATAIAGS Before attempting to calculate the risk oflarge events in the three areas under consideration, we should brierly address the nature of the statistical model to be used. It is usual to assume th4 catalogs such as these arc random, i.e., described by the simple Poi.+sonian distributien. This problem has received ample treatment in the literature (see, for example, Lomnitz,1966). In some caas the Poisson distribution has been shown to be a good description for large events, Epstein and Lomnitz (1966), and Gardner and Knopoff l (1974) have shown that the Southern California catalog, with aftershocks carefully l removed, is Poissonian. Other studies have indicated depattures from Poisson statistics (e.g., Aki.1956; Knopoff,1964; Shlien and Toksoz,1970). However, these ) departures are small, and may be disregarded for our present purposes. One graphic method of demonstrating the approximately Poissonian character of a sequence of earthquakes is to plot the interoccurrence times (Lomnitz,1966). In a purely Poisson process, the probability P that an interval of time T will contain at 8 least one event is given by P( T) = 1 - e-ra, l (g) l l I Here T3 is the mean return period for events in the sample. If the time between events in the sample is the variable t, then the frequency f distnbution of t is given by [ L s' I F(t) - e-u r., h, To 49, t f I '~~ ' ~ * ' ' * " ' " j J

r 768 M4cHAEL A. CHINNERY It is easy to show that the observed interoccurrence times are quite closely represented by equation (9). Figure 10 shows a plot of these interoccurrence times for the central Mississippi Valley catalog for events with intensity greater than or equal to V during the period 1900 to 1972. Clearly, the exponential distribution is a good description of the data. The anomalously large number of events at small interoccurrence times can be attributed primarily to the presence of aftershocks in the catalog. A similar plot for Southern New England data is shown in Figure 11. Data from the Southeastern U.S. were not available in a form that would permit a similar plot to be made, but this is probably not necessary. On the basis of Figures 10 and 11, we feel justified in using the Poisson model, and in particular equation (8), to calculate probabilities. In passing, Figures 10 and 11 make another point. It is easy to use the quantity mean return period of earthquakes in a sequence as ifit has a determuustic meaning. These figures are a reminder that the mean return period is er... rely a statistical so WSS155 Pet Vat.1.EY 19eo M97ll 75 + 84 EVENT 5 worke I aZ 5 RETUR># PEMtoo T, a G87 YEARS 20, 4 l T* g is-I r t a;w. yb T L I S i r i 0 8 2 3 4 3 a e NrERCCCUARENcE Tn8E (yeers) I l ' Fro.10. Interoccurrence umes using Nuttli'. (1974) data for the central Missinaippi Valley. The exponential curve would be espected for a Poissan dismbuuoo. l l quantity, and that its only real meaning is as one of the parameters describing the j probability distribution that corresponds to the catalog under consideration. Taz PRosaartITY or I ARoz EVENT 5 With the above model it is now possible to address the question posed in the introduction. In each of the three areas under consideration a large earthquake occurred shortly before the per.:ods of data that we have analyzed. Are these large earthquakes consistent with the later record of smaller events? Our procedure is simple. We take the linear relationa fitted to the frequency- { intensity data, extrapolate them to larger intensities, and make estimates of the 1 mean return periods of these larger intensities. We then use eg'4ation (8) to estimate i the probabili y that at least one of these larger events will occur in any 200-year t ll period, and specifically relate this to the 200-year period ending at the present time j (a 300-year period was chosen for New England, since the largest event occurred in the 1700's). l l l ,.... ~...... n - -...... _... mmund

I sElsMICITY COMPARISON-THREE RECIONs OF THE EA$ TERN U.S. 769 The results ar shown in tabular form in Table 5. We do not pretend that these numbers are very accurate. In fact, because of the subjectinty that ha3 to be used in obtaming the linear relr.tions [ equations (1) to t51], there is no way to make a realistic assessment of errors. We therefore view the numbers in Table 5 as being a qualitative indication of risk, rathe than quantitatis e. The results for the individual areas are discussed below. ,o SOUTHERN NEW ENGLANO 1860 -1959 32 EVENTS WITH I22 RETURN PER100 T = 3.13 ' REARS o d6 T, l s l-i W t -r/r i t4-Ae o i L 2 \\ f ! ! 1 E o 5 ic

s zo f

INTERCCCURRENCE TIME ( years ) Ftc.11. Interoccurrence tunes for Southern New England from the data of Smith (1962.1966). TAFLE5 Puon4arury cr Lance Evrars r3 Fora Rtcious or rur E4strax U.S. e l's. hen.hev af at 1.same One Evens Fe.asma Ow.1 Tim is.r..re in ta,rmd T r.n .we eno Iw ee r x e. avill alX 3X avlit glx gx Southesitem U S 19re 1 23 68 195 200 99 95 64 1969 2 33 117 417 200 99 82 08 Misatmppi Valley. IMO. 3 43 151 53* 200 99 73 31 19 4 Southern New England. 4 229 591 3467 300 73 29 8 1500-1959 g Boston-New Hampahtre. 5 371 1445 5623 300 55 19 5 15u%-1959 d 'r8 Y The earthquake catalog for the Southeastern U.S. described by Bollinger f1973) I is approximately 200 years long. Table 5 shows that, on the basis of the most recent i 70 years of this catalog (which may logically be expected to be the most complete at lower intensities), there is a substantial probability of the order of 50 per cent that j at least one earthquake of inten.51;y X or greater will occur in a 200-year period. We f conclude. therefore. that the Charle ston earthquake of 1856 tintensity X Bollinger, 19771 is entirely con 3i.4 tent with tne 1900 to 1969 data. b i a l

~. i I 770 MICHAEL A. CHINNERY j Without any question the larger.c earthquakes during the past 200 years in the l central 51ississippi Valley were the 1811 to 1812 New Stadrid events. Nuttli (1973) lists the maximum observed intensity during this sauence as X to XI, it New Stadrid,311ssouri; Gupta and Nuttli (1976) have recently revised this upward to XI to XII. Some question perhaps remains as to the validity of this value as a true epicentral intensity, since some amplification by the alluvium in the area might be expected. Table 5 lists the probability of an event of intensity X or greater durmg a 200-year period as being about one-third. The New Stadrid events were therefore reasonably consistent with the data for 1840 to 1969. If it could be shown that these were the largest events in the last 300 years in this area (which is not unlikely), or that the true epicentral intensity was somewhat less than X, it would be easy to increase the calculated probsbility to 50 per cent or more. The record of earthquakes for Southern New England is about 300 years long (Smith,1962,1966), During the period 1600 to 1959, Smith lista 3 events with intensity VII, and there are none any larger. Table 5 shows that there is a respectably high probability (about 75 per cent) that an earthquake of intensity VIII will occur somewhere in Sot.them New England in a 300-year period. The probability of such an event in the Boston.New Hampshire zone is about 50 per cent. The epicentral intensity of the 1755 Cape Anne earthquake is not well defined. Smith (1962) lists this event as intensity IX, which is probably somewhat high. The Earthquake 11: story of the L"nated States (NOAA publication 41 1,1973) lists this event as intensity VIII. Other unpublished studies hs.ve d9duced intensities close to VII. Whichever is correct. it cannot be said that this event is inconsistent with the subsequent seismic record. An equaly important result for the Southem New England regicn is that the i probooility of intensity IX and X events occurring within a 300-year period is quite } low. The absence of these events in the historical record is therefore again censistent t with the 1800 to 1959 data. Notice, too, tnat the return period for intensity VIII is l 229 years, which is consistent with the absence of such an event during the period 1800 to 1959. CoNcLtJSION We can make several conclu, pions from this study

1. The four frequency. intensity pl)ts that we have considered show a remarkable uniformity. All show a pronounced linearity, and have slopes which are consistent l

I with a value of about 0.57. This, in turn, corresponds to a magnitude b-value in the l range 0.9 to 1.0. This uniformity, and the fact that 0.57 is very close to slopes i observed in other areas of both Eastern and Western U.S., suggests that frequency-intensity data can usefully be applied in seismic risk analysis. In areas where data are poor or sparse, it would appear possible to combine data from as little as one intensity value with the apparently universal slope of about 0.57 to construct a local frequency. intensity relationship. Such a procedure may be more reliable than some of those in current use.

2. The uniformity of the shape of the t equency-intensity relation over regions ranging from the Boston New Hampshire zone and the Ramapo fault zone ( Aggar. val and Sykes.1978) to the whole of the continental U.S. suggests that the problem of nonuniformity of seismicity within a region is no impediment to the use of frequency-intensity statistics. The assumptions outlined in the introduction to this paper seem to be useful working hypotheses.-

I i .n

r SE!$MICITY CO>tPARI50N-THREE REGIONS OF THE E ASTERN l'.S. III

3. The question of the existence of upper bounds to maximum earthquake intensity <less than the scale maximum of XII) remains unanswered. There is no reason within the data them.selves to suggest that the three large events that we have considered are the largest that could occur in these regions. Similarly, there are no sta:istical arguments that a very large event could not occur in other areas Buch as Southern New England outside of the Boston New Hampshire zone) that have not recorded such an event. A rational, conservative approach to the estimation of the seismic risk at a site would include the possibility of events with intensity X or more an>vhere in the Eastern U.S. This topic will be discussed more fully elsewhere.

4. The validity of linear extrapolation of the frequency-intensity data has been tested by predicting the probability of occurrence of large earthquakes in the hist 'rical record, and comparing this probability with the known occurrence oflarge eennquakes in each of the three areas. The Charleston and Cape Anne earthquakes are both consistent with more ~ cent data from small events (calculated probabilities of these events are 50 per cent ore more). The New Madrid sequence is only slightly anomalous. The chance that such an event would occur dunng the past 200 years is about 30 per cent, but the chance that it would occur in a 300-year record approaches 50 percent. Thi o it appears that triear extrapelation of frequency intensity data to mtensities of IX and X is a valid procedure in these areas.

ACKNOW1.EDCM ENT Thzs rewarch was wppor'ed bv the Nuclear Rnulatory Commtwon. The author appreciates the helpful commenta on this pner received from O. W. Nuttii and G. A. BoJinger REFERENCES Aegarwal. Y. P and L R. Sykes (1979L Earthquakes. faults. and nuclear power plants in Southern New l York and No".hern New Jersey. Scienec 200.425-429 Akt K. (19ML Some problems in stattstical setsrnology. Zuut 8, 205-228. A:sermtssen. S T. t!%9L Seumse Ruk Studies in the Unsted States. Proc. World Coni. Earthquake o Eng. 4th. Santtano. Boilinger. G. A. *1973L Seismicity of the Southeastern United States, Bull Scum Soc. Am. 63, 1785-1%8. Bulltnger. G. A. fl977L Reinterpretation of the intensity data for the 1886 Charleston. South Carolina, earthquake, tn Studies Related to the Charleston. South Caroluxa. Earthquake of 1%6.-A Prettm. unary Report. U S. Geol Surtev P-ojess. Paper 1028.17-32. Chtanerv. M A. and R. G. North :975L The frequency of sery large earthquakes. Sctence 190,1197- !!98. Chtnnery. M. A. and D. A. Rodgers (19730 Earthquake stattstics in Southern New Englane. Earthquake Notes 44 s9-103. Cormil. C. A. and H. A. Merz i1975L Setsmic r.sk analyns of Boston. J Struct. Dw. ASCE !01,.a ST10. 272"-2943. Epstein. 8 and C. Lommtz (1966L A model for the occurrence of large earthquakes. Nature 211. 954-956 Gardner. J K and L Knopnff #1974L Is the sequence of ear 6 quakes m Southern Califorma. with aftersnocks removec. Pwoman?. Bull. Seam. Soc. Am 64.123-13% { Gupta. I. N and O W Nutth a 1974 Spatial attenuation of intensit:es for central l'.S earthquakes. Bull. 4 Serm. Soc A-r 66. 74 L751. ,I Gutenberg. B and C F. Ri+er i1956L Earthquake magmtude. intensity and acceleration. Bull. Serm. f Soc. Am, 46.105-145 Howell. B F, Jr. e 197D Emhquake hasard in the Eas:ern United States. Earth.if4neral Sci 42.41-45. Knopoff. L i1%4i The statot:0 of earthquakes in Southern Cahforma. Bull. Setam. Soc. Am. 54, 1371-g d73. L,mnitz, C. 41% Stattstical predicuon of earthwakes. Ret Geophys 4. 377-193. I .wta O W.

• 137.D The Mt-upni Vaucy earv.qupe* c.f 1811 and 1312: intensities ground motion and I

magnitudes. Bul; Sci a- &c. Am i3. 227-24A (t k t l l I r

r-- 772 MICHAEL A. CHI.NN ERY i;#s74 Stagratade reemence relatmn fur cenaa! %!>ueipp6 Vdey earthquakes. Soll. N atus. C W Ses m. Sec Am M.116s-1207 5hanal. A. F anc St N. Tcaz 4:.477: Eartnquake haurd :n Nr. Enziand. Sue,ce 194, : :.:71 shiten. 5. and N1 S. ToWz (1970s. A clustenng mods! for earthquane xcurrenen. Ball Sean Sx A m 60,17M '7=7 5 math. W. E. T i1W2i. F inhquakes cf Emern Canada and ad.iatent aren 15M-19f. Pabt. Som 06 Orr2aa 26,271-X;1. smuh. % E. T uML Earthquakes of Eutem Canada ar.d ad;acent areas, I?>19M. Pabl Dom Obt Ortsa a 32. s7-121. Venenano. D. < 1975v P-obabs.%tse and Statutscal Modeis ter seuma Ruk Ar:alym. %1 LT. Dept oi ( Cma Er.g. Pancauon R75 34 APPLIED S F.lawCLOGY GROLP I LLscots Lasonarcar, St.LT 42 Canze starrt Cru sa:oc t. ht tis 4CHtlSEm f/2142 51anuxr'pt recene4 Octooer 17, ;973 l l i l l l r b 4 l l l l 1 l l l l l l ---____,,.,v___m_.aw ,,,,,a.4

6 DW j w M e l1 b f l e a l l l l l

~ **= =.==. - w.. .y,._ * ..j = __e n..=,,..- ..,cy, -sg .. ~ .-e,. cc,

p. e......~>

o 71 , r e, .h = ~e ..~.. .s. . y-.. ~ - +.. .t.- . 'x. .v e.. ~... -.. e,..* s .. ;s' 4.

e.. - s,2**....

.r. .-e...,. .-. < c n.. A. r... -* .. L.m - y . % r *'*,..+ r 4;- s. ;... t. .w, yk_ c.n -. o,.. .n f ,~. 6.~,,,s.g...,. a a a-z. .. ~.. -4...,.:... ,..,-~,c-..n .n s ; t* .y,, s.,, s. .a - g , r, -r, wf,. 3 u ..a .m y -.a m.. -,,,. o.,..,j v. y. ..2,,.,..-y. :.m. ...e... .r;,

u. p. t,.

.s.. r.m, .s -.. ?,. e-2 a , ya.%.s n. +. f., m, y. . - s* w'.,,' e 4. '-.

• c.

.,...,= 2 ,a....... c a.~~:~. ". u '. g:.,%i.W:-.w'- 9r c~ . s: .e. .e'....

• ...e 4,m i*<.t.1 e

,1-a.. W%m.-.. :e ; ;..,; 2 ,N, k..:. " x.), Q' <% c y.-.-g y ; . * ~.

1. .,,ny e'-

. < ' - *.a s -f .. ~i.' s. ~-- /na .,f.'t,....

' f,J L: l,.:..

..?' *,,. f.* ',.' ;".; ' w.,

• , '.9^
• f[f.,,,

y?n.rg.?;..'+q **. ~ T;.,-f .%. A % ~.. '*c.:... A,R 'if-*,: ':*v:lJ e c> '. * :. * ^ e ~*Y..a.f3 .~ m;..c->.' e..' .;gw r.s s.-,:v..*. s. - -v .".*t . - ~ * ~ *, a/ .. *,c. se. -..m...., n. : 4,. ,,,. g g,... Q. . e.,. mq. Q. s ..,;+a .cs w.s.. a t c. 3,.2._.~* }s q. ..e ... g $fYY *, 2y~ G :i -t .q. ; ; e '..

,y --n g,,; s

n,e m

• * $'b ;G.N; *y

-.s., 5 G. f, p G-h r ? ~ W "* * ^ * * ** 3 s =~?.

• -. ~

s E %'~ '. s * ;*' ~ p.... .s.. -~.u.. ,-. - v , C q.-ww g~; - ~ * ~,. ' ~. 6.,,':'N ~ ~* m-i. n.,~ t ,...Y... C, ~. ~,:;.. *., j, :.... ..,4,.'. " *..?. e ; ..y9 f,, c.I9,3 5 c.-..*.4., - J ..)- .+ v_ .~; Q. a%.p..

e n,.c.f:.., !..',,, 9- !.,i '.r ?...

. ; ~, 4 r rt.y...* ~=.C a A.g.4..;.Ayte. WQ. ...,..~.'.._.i.g.p,....,. <. q.. " L*. %. -v,.,,; ;,....,,n....~.... i '.w:.'...,.. .. '.m 6 2

, ; %.,3. gg ", 3,,, g,..,..z..

p. p-y,.,.~,.

.%,m. g,.,.a..g -r . s x. s..y. +: .~

s..:.

.s .w s %.. C, c . g.r r ..- v y. r' w.n eq. 7,,..,... .1.. s-7 .o. .. c - yy ...? c -t'- .y;g.g. ;) g a g 34 --L*. ..g; g,,,..h. tg.,y, e ?~**: h. n n u.a.,;'. Re,J ott.;,7w ? w". n :w w..,.%*w::M m'M%m..e A m-s .c,-..c.e=.>,~..

w. ~, r m w" qty.w.w a

r..

....t m.,. m;it.%.W. _S... M...M..w:. n@...n..shM.5. w.g_. % _. W.-MW._ igg. q.t - -w. ~ M. ms .,.p.-,...- .g...

s... a*.**_, C. ~

..'. ~T,a--e.- <>,- e.. e..~ 4. n,..... ; w,.g,..n.... - < - 'A '?;. Q, ! ;* O *l h '.:.2 - Y.245 ?.N ? W &.M**** &#5E.N M '5 G%.'U 4 ' - I'Y ~ # $WF.~~A'*' U*Y.*Q&'.- = .y .,....m,o.m,~.* .,.,9.J..r.j -.,.. -.. s. n a .... m,,.... . ~ s ... s...:.. < P,m,...d, .r,,..,. .= e.. ~. w..~ ~.m.,o : v m.#y.,. f,.. _.a.. p mn tv . s...~e..vs. 2 .g,g m. m. .-.4 e 9:q.,.,gy.u. n.mf. ~ #g.w.. %..,p m.,-,.m.;g.,,.wp ..m. ~. w .m w wg w::.gm..7,.-m. a.m.eg a. s .,g.., 4.. ~c. - w ~... m-. .a n.* 4:.. . w.... w..n.a. .a_. P: o. cn &.r o w.

• A cv'<::. : >. vw.n.4 "*.< -

"'*J.4.'1".'ss*T p!,'d.* M

s..n :r..

.' g. -e.7 s.. . r' ..w : . %.m.e -.h.G.; 3,;. . J. 4,5g;q@44..-;,g geg;g w <: M <-.Q t r , J J, .~.. @ g* .~..;...a.6-= 5 4 -L - p 3 -- .%..i i nInvestigations of'the Seismological (Input y. g;w-.u%,.io tNNafetElielig@o"f uclEfr$dNe$" M$%@MkMNh3 WM W W' 95249#.W9M W W E #3 Mi45$:We d@q$cmu:/vWn;s Sm% A.0hm.,e.n,%. % Reactor &s1n hew EnglandM.mn M.. g : 3.n

.2 r-s

3. =M.

.AG.

p. vw;.,.. w..

e,*.,g..a. 9, pu.. w... m,. G - s - ,.t.-'a '. a vt -. e -.cc. n.- ..a q,- , L.

.. u..,.

,Z 7 ;..A,.,:., 2.r e .... '3.N,, a.w*. ~ - <m.q m.. + - tw ,. ~t. t ........,. -< e.t.. s. ~,,c- . t..s.,.,9.,,.A?.. u..<,- c,. '. Aw.. r2.g-

2.,. W,..,._.a..

.'..'.s.:y... A;im. -. '..... u U$[ l .d..,..,cC,i>m&,. + .... +...

.e

?.. ... 6,,T.W. -Ws. -.,,..,... _. m...:. s ..-{ $ '.. a .-.,,....... :,:3,e m .m~.. .. Y:.

• A..

..,g, s. ;r-.._,..,dr,.... z . a. s:.;......,,,... .-.a,n%$,. n

D:',;

f. x ~; . w m.. c -c.m.M e - ~ %' 5..... a ',. ~.., epa:el b t.n e.w e,. ear iienlatiry Comm... k'f:n.., %,., ..'s %., f. m:&..,.f..*'.h.Y'f. %..p=sN Q. *:- Q ': ' i x,'..r. -~ " Pr% us A.- : [k.(.... e.

.Mdk

.a -N fk'

• • C EU Rd-0 1(. 5h. 8kh.......k.

j r '*f. .~. e . Lm, co_ n.i La 3 oratory [ 4 'j"'", N, w !* *q.: o e f ~. - = x s b g1 e d. r M.W...A C' *.i r. ~..c .T..c '. '. ~. "e 's O F *. r '~.O O L s' CY 9 b .1 ^ k. Ltmcm. Ws s acs ms m'3 l .g, . r. - , { v., ~ 1' - s.f e 4- -. - si** l .s . h.. ...w '.. g ....'r ,.= ~ ,.n ^ 1 -, ,y (. II w Mf. t I g,ge w *.. me ), ~V. m 9 E.,- 9 N pe s M g . *%IhW$ g. S: l .:. =a..... g w. =.,~... +4 g .. z u .. x.. g -

w....

c.,. =. x. a., e.a. ~..

.~c..>.. ~. ~,. ... m. ( .h. g.-+# .F .t.,,, e,'e ... =. ,.e, -f,, 7.,.**.* -e g. ij.x..n - *,.y g..,,-. l

• , g_

3 J *."

• • • '.i,T,,

.,f 2.a;.. i' . L _ a, M'Q,.. - b' a.

tassachuse tts In-titutt' of T chnology Lincoln Laborawry AN INVESTIGATION OF MAETUM POSSIBLE EARTHQUAKES A:: cal Report l Project

Title:

Invest'gations of the SeismologicalInput to the Safety Design of Nuclear Power Reactors in New England. NRC Coattact: NRC 04-77-017 Prine: pal Investigater: Michael A. Chinnery, Group Leader A;; lied Seismo:ogy Group L :ccIn Laboratory, MIT 42 Carleton Street Cambridge, M A 02142 Per:od of Centract: 1 January 197 - 31 Cecember 1977 I' Aqust 1973

Abs *ra

• This report describes resear:h O'arried out under :TEC Contract "?C-C.' *6...' y'

.i.. d. e- *. ".. a. - a. *

• s.d.' v' a....-.' 9 ~4
• O 2.'

u' a..- a...k. a..-.' 9", 1 a'

• ^

~ y.. s w 4.. a 4. .s..,4.../ .,a,.a4 a..,e e 4.,.,., 34...a....e .,. e 4 g .w...e,. s. 4_.a. 4.c n of =axi=un possible earthquakes sh0vs that all available =eth:ds are e=pirical and lack a scund physical basis. 291dence that even the e=pirical =ethods are valid is very weak, pri=arily because of the short 1ength of the earthquake re:Ord in =ost areas. An atte=pt to use g10bal earthquake :stalegs to exanine the regional variation cf =ax1=un ;cssible earthquakes is unsue:essful. It is de=enstrated that saturatien of the s.~y 4 s....e.a.-.,4.,,4..,...e .s.. ._2. 4...a.e .2, ale a 4 w4a.ses 4n.. 2....a , w 3... to =ake a values for large earthquakes very unreliable, and to obr:ure o the presence or absence of =ax' nun possible earthquakes. A progress a repcrt en a study of : lev Ingland crust and upper tantle structure is 4. c.,.a

a..

4 6 e ) J l l l l 1 444 l e

Table of Centents 111 Abstract Introductien 1 1. Maxi== Pcssible Earthquakes: Current Status 2 1.1 Introduction 2 1.2 Definitions 3 C 1.3 Approaches to the Proble 5 1.k Physical Argments 7 c 1.5 Arguments Using Iarthquake Statistics 10 1.6 Use of the Level of Seisnic Activity 12 a 1.7 Pattern Recognition Apprcaches 16 1.8 Other Studies 18 1.9 Discussion and Cenelusiens 19 2. Analysis of Global Catalogs 23 2.1 Characteristics of G1chal Catalogs 23 l l 2.2 Iarthquake Statistics 2h 2.j Saturation of the Magnitude Scale 28 2.h 2.e ISC Catalog 32 2.5 Events in the Aleutians-Kuriles Region ho i 2.6 Interpretation h9 2.7 Dir assion 57 2.8 Conclusions 60 References 61 Appendix: Progress Report: IIew England Crust and Upper Mantle Structure 68 iv . ~. ~. ... = - _ _ _ _

i I i 1 I i 1 4 4 5 i i Introduction 1 This report describes research carried cut under NRC Contract NRC-i CL-77-C19 during the period 1 January, 1977 to 31 Dece=ber, 1977. The i =ajor effort during this period consisted of tvo studies ai=ed at evaluating l* the possibility of esti=ating the naxi=u possible earthquake that might i be expected within a given region. The first study censisted of a review and assess =ent of available scientific literature en this topic. Since =uch of the research in this area has been carried cut in the Soviet Union, this review provides a reasonably ec=prehensive set of references, sad a discussion of the f l varicus approaches which have teen tried. i 1 The second study was an atee=pt to icek for evidence of upper i l bounds to earthquake si:e within global body vave =agnitude catalogs, i and in particular in n e ISC catalog. This study soon turned into an t atte=pt to understand the sources of bias in the =agnitudes listed in this estaleg, since until these are understood it is i=possible to i search for =axt=un possible events. It transpires that these biases, 4 1 4 together with saturation of the =bscale, =ake =b catalogs essentially l useless for this type of study. l A third area of research, into the crust and upper mantle structure of New England, got undervay during the period covered by this report, and a progress report is included in the Appendix. 0 l l l

2 1. M.TJ:UM POSS!3LE EAR"'%UA.GS: CUP 2E'IT S'2ATUS 1.1 Introduction Ve veuld like to kncv vhether or not there is a limit or " upper bound" to the size of earthquakes for a variety of reascas. First, earthquake size is usually intended to be a =easure of energy release. However, energy usually varies strongly with size. For exa=ple, the standard relation between =agnitude M and energy 2 (in ergs) is log I = a +bM (1.1) c o o Bath (1966) reviews several estimates for the constants a and b, and n shows that b, appears to lie in the range 1.k to 2.C. Since the number N of earthquakes is usually described by the relation log N = a - bM (1.2) where b is about 1 (see, for exa=ple, Richter 1958), the :otal seismic energy release is decinated by the largest events. We s'2all have reason to questien both equations 1.1 and 1.2 later in this report, but the eccelusien appears to re=ain valid. Analysis of the energy budget of the earth requires knowledge of the rate of occurrence and energy release in the largest events that oe:ur. Second, 3 rune (1968) has shewn how the relative slip of tectonic plates can be esti=ated *:ms ea~.hquake M e. and shoved that the total slip is de=inated by the largest events that occur. The fundamental question of hev much tectonic :otion is released in seis=le slip (Davies and 3 rune, 1971) can only be answered clearly once ve understand these large events. And, thirdly, the esti=ation of =aximu= earthquake size is impor-tant in the esti=ation of seiscie risk. The possibility that large events =ay occur, even infrequently, in an area can lead to a seismic M _K

) i k I 3 i nazard that is unacceptable for certain critical facilities such as l nuclear power plants. The !!EC Rules and Regulations, Part 100, Appendix l A, set out the seisnic safety standards for these structures, and define the Safe Shutdevn Earthquake to be based en an evaluation of the "=axi=u: 4 l earthquake potential" of an area (Hofmann,1974). The purpose of the t j present study is to assess our ability to esti= ate this quantity. We can usefully divide the overall proble: into two parts. First, f* vhat is the evidence that earthquakes considered.as a global phenc=enon i have a =axi=u: pessible si:e? And second, how does this =axi=u: possible i size vary ft = region to regien? The first question ought to be much 4 simpler :: ansver than the secend, and it is icgical to exze.ine it first. Hevever, as we shall see, it is difficult to give convincing ansvers to either of these questions. 1.2 Cefinitions t There are two i=portant definitions that we =ust explore before ve continue. The first is the definition of "=aximus", and the second is the definition of " size". I 2e te: :1 "=axi=u=" is not, unfprtunately, always used with the sa=e meaning. Cne definition is the obvious one, which refers to the largest possible event that can cecur given the physical conditions of the source area. A second definition, sc=etimes used, includes the concept i l of pr0bability. A certain probability level =sy be accepted as being i " negligible", sceerding to engineering design standsrds or other argn=ents, t I and the "=axt=u=" earthquake defined as one which vill cecur with this probability level (or less) during the projected lifeti=e of a structure. These two definitions are very different, and it is essential that i they be clearly distinguished frc= cne another. We shall use the termino 1cgy 7

h M for the "true" =axi=um possible =agnitude (E for the =axi=u:2 =ax =ax v possible energy, etc), and M for the =agnitude that occurs with probability F, which defines the accepted probability of "negligibility". As we shall see in the next section, very different methods =ust be used in the esti=ation of M and M =ax =ax The definition of earthquake " size" is even more difficult. There are a large n=ber of quantities which atte=pt to =easure this size. A partial list includes: a) Body vave magnitude (=b) b) Surface wave magnitude (M ) s c) 100 second period =agnitude d) seismic =c=ent (M ) e) radiated seismic energy f) elastic potential energy release g) naxi== epicentral intensity (I) h) =aximu= epicentral acceleration

1) local =agnitude(g)

The basic proble=s here are not only to decide which of these measures ~ of size are the most appropriate for a given situation, but to recognize t l that the relationships between these measures are in general poorly understood and in so=e cases demonstrably very non-linear. In parti-I cular, sc=e of these quantities have built-is upper bounds which can obscure the search for a funda= ental upper limit to earthquake size. We shall exa=ine this proble= in = ore detail in section 2. ( An additional ec= plication, which arises in the literature very frequently, is that the ter= =agnitude is so often used without proper l definition. All practical measttres of magnitude are restricted to some 1

2 =* eJ limited portion of the seismic spectru=, and are closely tied to the =ethc2 of measure =ent e= ployed. ~here is so =uch variability in both of these facters that the te:: =agnitude alone is al=ost =eaningless, particularly when the characteristics of large earthquakec are concerned.

,uite often, in reference to the lccal seis=icity of area, the ter

=agnitu e refers to local =agnitude.%. Of all =easu es of =agnitude d a i +.4s _s _,,,. o.c..... h. 4.s. .,,,,.a 4.# 4 n .w v. ..a s .4,, +,. a.tc,. a o 4.,4 n.

• 1..

%,y ....w ......s Richter, and designed fer local shocks in California. Its definition is r va..., a-bi. =.~s, a.n.d. a. f e.-s

• ,,o
• .k.e

.' e. e a-4

• w o.* *,.> a. ax_-",.

.-a.

c. d.a. d 3

trace a=plitude of a specific instru=ent ('Jood-Andersen seis=0g aph) at 4, a specific distance (100 1:1). 3ecause the instru=ent vill reccrd a vide range cf frequencies in the short period band, and because there is no seis=ic phase identification, the significance of the =aximum trace a=plitude is not clear. For s=all earthquakes, the =axi=u: trace a=plitude i vill often refer to body wave arrivals t short distances. For large I ) earthquakes, the =axi=um trace a=plitude vill usually be associated. vith fun a ental mode or higher =cde (L, phase) surface waves. d i t a The principal usefulness of 1 is, of course, that it is a = essure u e ,s, #. e. ,..a. ...,.4 n.4a.w. .a,

1. .4 u

.e _ a. a ,.a .. r.,.,,,..,w,....,, 4 e s + ha. a,., l relevant to engineering considerations. I= prove =ents in the esti=ation' of R (Kana=cri and Jen.ings, 1973) =ay lead to a =cre consistent scale, a u but its relation to far field =agnitude determinatiens is still unclear. + 1 1.3 Acerca.ches to the prcble: f .._..._s,,,. ..,,s.'..*.'..a. y.. .' '.. a r a -e -.'..= '. c.'., e='y. *o.-a.+

• ^

'..e v. heart of the proble: cf the esti=ation of the -wi=u= possible earth-i i quake is quite small. The majority of these are the work of scientists 4 l in the USSR, where there has been a 1:ng ter: interest in this topic. i ? 4 5 t w*-+v-'*- ---r--'+"-v'y ~~ r-+,v' -'ec= r-

• e-

'"-w'. "e

• - < + ' " " ~ ~ - - - " ' " * ' - * ' ' - " -

" ~ ' '

6 unf;rtunately s:..e of these papers are hard to obtain and difficult to read. A nu=ber of apprcaches to the proble: have been propcsed (see, for-exs:ple, Shenkova and F.arnik, 1971.). First, there are a nu=ber of broad argu=ents that atte=pt to li=it the upper site of earthquakes en the. basis of physical principles, including fault gec=etry and slip, and the strength of earth =aterials. Generally speaking, these arguments =ake a convincing case in favor of a global upper bcund, but give little indication where this =ight te. A second apprcach uses earthquake statistics, either in the fem of frequency-=agnitude data er =cdelled by the theory of extre=es. These two analytical techniques generally lead to si=ilar results, but both turn out to be severely li=ited by the definitions of =agnitude used. A third approach, which see=s very logical yet which lacks any convincing physical basis, attempts to relate the size of the =axi=u= possible earthquake to the level of seis=le activity in a region. It vculd be very nice if such a relationship vere to exist, but there is no clear evidence that it does. More recent approaches have tended to focus on information frc= non-seismic sources, such as geological and gec=orphological data. Sc=e of these' approaches are statistical, using pattern reccgnition techniques. Others are =cre deter =inistic, and atte=pt tc link long ter= geological fault =cve=ent to short ter= earthquake i slip. In virtually all of these a.proaches cne proble= predo=inates. The reccrd of earthquakes is relatively short in = cst parts of the world. ':ata before abcut 1900 are generally qualitative and hard to interpret. Adequate seis=ic networks have only been available since the early 1960's, and (as ve shall see in section 2) there are still proble=s in w _.=m._ g, M-

1 ...g ,4. .s r.'p

o. g. u.,3 3 i..a

'.... u..... r.n.. u.ge.n .g A.e.v 4. 4 .. u.., ... f g. 2 4.r.#.4.., ' '. '. O e S *. a '..'.#5.*.***#..**.*-a'

• • • a#..'.

1X.. - y'"e " a'.# k. ' e

a. r

• k.a ". * '.<.
• S a

.v. y 4.. 5 g., 4..r 4.,... s. 4.,.., 3... .w.g. 4 s. ......a ..], ...a.ye ....... 4.3.43 w . 3.,... a... .. a ..~f.. .. -.maa y... . '.. a. e *.* #.".i.*.* '.

• S *. #

1 *. *. S, .#5 ".#.."*.6..f ...3 r.. p..w n.. .s y.

4..c g a-4 s...

3 oy 4.. .w. . '.4 2 4... ,p .a.v "."^. y* ^,,, S e d..y y."*. a '.. e S. .y .. y O. w. f S. a7. 4 ., ~..-....a e

• • v,. a.aa* C *.

^# ~'** ^# '. e "."..# '.' *.. S a.' . s..

• S *. *
• S **

.a. e.. e... *

• 3 *.

a... .e e 4.. 4 4..e, 4-3 p. %..g. g.

n. 5 w,4..g 3.

s..-.. 4..

...4- .m.-.+-.... a 37 a .e . e.4 . a..,, 5,.,..X.e.+ <'e. S. 7.. es,S*.4 .S 7.7 14 7.2 s.,. 4.,. S.,. s. ... 4..g .wa..S..,,7A -y.7.. a 4.~ we.w % e. 3.4. 4 e..,, 4 a. e. 4 4. e.7.. 1E.. v. A, 1 -4. .C. w... a... . * ~ s ....... s 3 r-(.s e. ., a a. . S..k. 7.... k, ...~. 4 e4,4.e

7... e.e. 0.e

.., 4.e e,, n4. .4 a . a. -....... w y.. ',apf.7 '). 'e.g y.., so.w .e s.'3, oro 4 g... .,. c. 7..

4. re.. A

m. *. 'g.-.

y. .....s .s. y . e 4... a.. S, s,.w. ,. _. e 4.S. .C w.. #. 7.. ....y. .S .C

4. S

.e.. 4.4 4.., .w.,44 S C.. S a-4. -c .e.4..'.........y y..

• ...44

.L wC,.AS. 'Je,ea '4,. A

a... w 7,.. % / - Q'.;'.7 ). 1,'.. h.a.

._ S w,l.a4 S..:. w s, ..v .. 1 4 *r. 7.,.,

s. ~-

4. ... a-2 A. y > 4.... .w. V. > 0, 4...... -. 4 .w. ,....e. .a I i u.aa.4 - .r.e.,.. '.,,.. S e. S. 4 a.. S. u

e. aw..,,,

a 4.a. . s.e.,. 4 -.a . u.n.,.u.. u e a g. "e *.*.a '.'.' e *. .'..a *. V.. "g.""'" a.' s .*.* e S a' k

4..e v

.4-

5..

e.,, r.. 17 e *

• e~~...# '. ". '. *.

.V. o a a s 3, 4 e. r.,.w... a 4... A. s... .w S

4. a.

2.. 1. .w. .L....a 1... .s.... .e f ..-\\./. ~a. w,,... '4e S o r. it ,. g...... e w.,. a (a,. 4.. ..5 s,e.a >a s.. .yy .44 %.a ,.44..S,..>e._.ue w.... %... e. .~ %..e... ;. A. ayy. n7 4. 5 n,,.e. mn. u.. e. -. s.. 4... .s - 7... a n. I ..2.:.,.... .4 i.. w.r. w,... e .or.

a,.S.e.... a,.

4. a-. 4 4 S

-. 4

.a-4.. 4

4.,,,.

4- .g ..~ .a g. .a a,. ?.- &. A. v...,..(-~~',,, - QM, i ...,.,.a.. a %..,~. a.., 3. S.. s.... '.f, e -.y ... y y.. ...s-

4....,,.g.-

4..., 4. 4. a-4....w. w-.- ........w.

s..=. S. a~

#. 1'y.'"*S'****y*.#.".. .r .a..., 5ei e 4-. ". ...,.,.,. %.n... A. X.4., a A. " -.... e a. 4. S..e.,, u.. ,., 7 . u.a. S,, u. a .aoa a..,

• S *. *h. *..~. *.# '

1. . ^ * #. a" k.
2. ..~.

• %.#S

~.. ...g..,.i .. 4 4.,.4... - ... 4 . 5. a. .a ...e 2.. ~..

8 area (see also Esteva, 1969; Veneciano, 1975). It seems likely that this study reflects a general belief that areas of lov seisticity should have lov upper bounds to earthquake sice-(see section 1.6). It is possible to go sc=ewhat beyond intuition. Tsuboi (1956) has proposed an upper bound to earthquake energy. He first relates earthquake energy to the volu=e V the strained region around the source, then assu=es that the strain is unifor= throughout this volume, and then uses field evidence for the =aximum strain which the earth's crust can withstand (about 10' ). Then, if V is li=ited by the thickness of the crust, an upper bound to energy of about 5 x 10~h ergs is obtained. It is hard to c assess the validity of the assu=ptions used in obtaining this result. A very si=ilar approach has been given by Shebalin (1970), though it is less convincing. He quotes linear relations between earthquake =agnitude and both mean length of focus and vertical extent of focus, from an earlier paper (Shebalin,1971). He then uses li=itations on i i t both length and depth to set an upper bound to =agnitude. The validity of his starting relations is very much open to question. Similar procedures have been outlined by Hofmann (197h), who describes [ hov =agnitude fault-length relationships (e.g. Sonilla and and 3uchanan, 1970) =ay be used to assign =ax1=u= =agnitudes. Obviously this type of approach presupposes that we can clearly define the location and length l of all active faults in an area, that breakage beyond the present fault length is i=possible, and that the =agnitude-fault length relation is l single valued (this is equivalent to proposing that all earthquakes have t j the sa=e stress drop). Each of these assu=ptions is difficult to justify. l Shenkova and Karnik (197h) rr.ise the possibility that the rate of strain accu =ulation =ay set limits on the =axi=u= energy released in an t

s 9 earthquake. " hey indicate, for exs=ple, that if upper and lover bounds can be placed on a Benioff strain release graph, the =axi=um possible earthquake vill be specified. ~his approach is meaningless unless the record of earthquakes already contains at least one =axi=um possible event. These studies are typical of those atte=pting to use physical i argn=ents.

• he strength of rock, under various physical conditions, is not well kncun. However, we kncv even less abou*
• 'a

'4 '*ations on the size of the :cne of slip, and it is this variable which probably li=its the.usefulness of physical argn=ents. The largest kncvn fault area is prcbably the 1960 Chile earthquake, which vas about 1000 k= long and perhaps 200 k= wide en a shallev dipping fault plane (Kana=ori and Cipar, 197h). There do not see= to be any convincing argn=ents why fault breaks could not be larger than this en occasion. Could the j entire Aleutian are system break at once, for example? l The effect of strength of rock is related *o stress drop. The basic proble: can then be for=ulated as follows: Seistic =c=ent M is O defined by . (1.3) l M = uL'a*D t O l vhere u is the rigidity, L is the length (long hori:cntal di=ension), i '4 is the width (shcrter vertical or icvn dip di=ensien), and D is the average fault offset. he stress drcp 13 can be -Titien ac = n u3-(l.h) M where q i:. a gec=etrical facter which typically ranges frc= 0.25 (for 1cng strike slip faults)-to 0.75 (for long dip slip faults), as is shown ~ by Chinnery (1967). f -,--v,

10 So ve =ay generally write M - 2L'4 a0 (1.5) If stress drops are roughly the sa=e (about 50 bars) for all earthquakes,' as has been suggested (Kana=cri and Anderson, 1975), then li=itations to seis=ic =ccent M depend only on li=itations to the di=ensions of the o fault area. Ecvever, questions about the constancy of to re=ain. Sc=e studies appear to indicate local stress drops as high as several kilobars (Archas-beau, 1976). In the eastern US, the occurrence of =oderate sized earthquakes in the lever crust with no surface expression of =cve=ent veuld appear to require rather s-* faul-Ad--asions and correspondingly large stress drops. To take an exa=ple, if a fault area 20 x 20 k= vere possible in an area of stress concentration i= the Eastern US, with a stress drop of one kilobar, equation 1.5 gives a seis=ic =cment of ever ( 10 dyre-c= (equivalent to an M of over 7.5, see Figure 4). This is s probably larger tha any earthquakes so far observed in this area. 'Je conclude, then, that while physical argn=ents support the idea l that there =ust be an upper bound to earthquake size, and suggest that there ray be a substan** ='==gional variation of this upper bound, we cannet yet ccnstrain the appropriate para =eters enough to esti= ate the l i sizes of these utper bounds. r i 1.5 Arzu=ents Usin: Ear-hcuske Statistics A variety of authors have atte=pted to use the statistical char-acteristics of the ea-thquake record to estimate =axi=us pcssible earth-l quakes. It is not at all clear that existing earthqua'<a catalcis are good enough for this type of stu:iy. Certainly,.in the exa:ple C.iscussed j in detail in section 2 of this repcrt, it is clear that proble=, of l e t

saturation of the =agnitude scale and individual station detection ec=pletely cbscure the presence or sesence of upper bounds. There are two possible apprcacnes to the analysis of earthquake c atalogs. The first involves the use of the frequency-magnitude curve, which is discussed extensively in section 2. The other is based on Gu=tel's (1938) Theory of Extre=es. Gu= bel described three asy= ptotic distributiens which =ay be used to =odel the distribution of largest events occurring in a sequence of equal ti=e periods through the earthquake record. The Type I asy= ptotic distribution of 1*rgest values corresponds to a linear frequency-=agnitude relation, with no upper bound. The ?/pe !! asy=ptetic distribution includes the case where large events are less frequent than vculd be expected en the basis of s= aller events, i.e. a non-linear frequency-=agnitude curve. The Type III asy= ptotic distribution specifically includes an upper bound. Algebraic details can be found, i. for exa=ple, in Yegulalp and Kuo (197L). i Applications of the Type I distributien generally acec=plish nc ore than the use of linear frequency =agnitude statistics, ard no upper bound is included. Papers using this distribution include Epstein and icnnite (1966), Gayskiy and Katek (1965), yllne and Davenport (1968), Connell (1963), Karnik and Hubnerova (1968, 1970), Yegulalp and Kuo (197k), Shenkova and Karnik (197h) and Shakal and T0ksoz (1977). Though sc=e Of these papers =ention =ax --- -=gnitude earthquakes, it is clear i e that wh:t is discussed is the cuality :(=ax, the =agnitude which has a probability of occurrence (during sc=e fixed period) that is less than o. Studies that atte=pt to use the Type !!! asy= ptotic distribution are pctentially =cre interesting. These include P'ei-shan and Lin I I i J

12 (1973), and ';egulalp and Kuo (197h). The first of these studies does not define the =agnitude used, while the second is based on Gutenberg and Richter's (195b) data. They can both be shown to be formally equivalent to trying to fit the frequency-=agnitude curve with a truncated distribution (Cosentino eji al,, 1976, 1977). We note that Knopoff and Kagan (1977) have argued that frequency-=agnitude statistics are to be preferred over extre=al statistics since the first uses all of the available data. To anticipate section 2, there is no doubt that saturation of the M scale begins in the ran6e 7-7 5 It is interesting to note that = cst s of the esti=ates of M frc= these studies are greater than M = 7.5, =ax s and the vast =ajcrity are greater than M = 8.0. As 1cag as saturatics s of the magnitude scale is not considered, there is no way that the results can be unsnbiguously interpreted as indicating the presence of an upper bound with regional variations. 1.6 Use of the level of Seismic Activity Perhaps tha = cst persistent atte= pts to study the nature of earth-quake upper bounds have been =ade in the USSR by Ri:nichenko and his co-verkers, beginning with Ri:nichenko (1962, 196ha, 196hb). Many associated references are listed by Ri:nichenko and 3agdasarova (1975). Ri:nichenko's basic postulate is that there is a clear cut upper bound to the energy released in an earthquake. Setting the total energy C 2 release E = 10 jcules, he discusses the eroblem in ter=s of I and cax K_ax. He uses an i= plied relationship between energy and the observed quantit y, =agnitude, cf the for= log I = a + bM (1.6) The particule. values of a and b used are not quoted (and are still open particular definition of =agnitude M is net given. to question), a-d

• ba i

I

.- ~ I J 11a i .. -: e,.,,, g,.4., a. ,.,.w, ."a'. **. vas 44'^ c"'+ o- *..,.oss'ble e.4 4 w s. ...e to deter =ine K directly frc= the observed earthquake catalog of an =ax a ea. He has therefore focussed on the possibility of establishing a relationshi between I and the level of seis=ic activity A in the nax frequency-energy relation i i .c g.i = n - n.. y.. ) ..,. ) i C (A is therefore the activity at the reference energy level K ). He has O discussed the for= cf the relationshin. A(Km ) in several pa.cers.(Rizni-chenko 196ka, R:nichenko and 3agdasarova 1976 and others). Briefly, his argn=ent is to relate the energy K Of an earthquake to a volume radius R (fer entral Asia he obtained R' = 0 315 10"'1C), to average the activity Y u a A over a circular region of radius R o obtain I, and then determine an For Central Asia he determined e=pirical relation between A and K nax (Ri:nichenke and 3agdasarova, 19761 i icg I = 2.Sh + 0.21 (K -15) (1.8) nax .. -. n.4,, < - u,a a,. s, ,..a a e..e ,4.....n. w s \\ l i log A = 2.Sh + 0.39 (K -15) (1 9) nax i n er ( 2 ese equations are intended Oc be valid for 15<K<19, or 10'2<I<10 # ergs. i ~'he fer of these equations was ierived ver/ artificially (Ri:ni-chenko, 196ha). K vas si= ply chosen as the-largest event for a given naX ,. 4. c,.. r.,.. e n, s '.... a.- 5.. ** .a. s a_. ' a. ), " A. a' d a. t a. _'.a_ d. fm-. +5.a. s.. - region. The 100 cf A assinst K.,x had 20nsiderable scatter, and a + ,4.,,,....._. ..,1.s. n..as ,.4.,a.o.u.. .., s

v..

v,.,.,a,. s

n.. s.. <

u..,. G... a. ... nax. ~akhareva, 1971). In 196h the constants esti=ated in equation 1.8 vere l l 2.30 and 0.20, so there has been li :le change in the relation in the i I subsequen: 12 years. The difference in the siepe found for Japan (0 39 i i 44..,.

• 4e.$.

4. A. ^ d.- E 4 4, e ..sq.4 %.. N te eh. .e) er - w V. d

lh i ( -05, log N = A - yK i I I K K = ic, r e log E=ax Fig. 1. Ri:nicher2.0 postulates a clear-cut upper bound to total earthquake energ I, and assumes a linear frequency-enery relation for energ values belev I =ax \\

15 Obviously, the proble= in this approach is that K needs to be =ax determined in sc=e regions before the general law can be established.. We =ust allow, hcVever, the possibility that successive application of the equation in various regions (e.g. Gorbunova, 1969: Druuya and Stepanenko, 1972) =ay i= prove the constants by an iterative or " toot-strapping" =ethod. The icgical basis for the expre::irr. 1.8 is not established. Whether or not it verks in practice is less clear. The authors ec= pare 31 large earthquakes in Japan with the predictions of equation 1.9 Twenty-one are found to be in agree =ent, 10 are found to be larger than the predicted K=ax, though the authors note that uncertainties in many of the epicenters =ake it hard t: =ake a fir = conclusion from this result. The situation is far fro = satisfactory. The existence of a relation between K_ and A is not proven, and appears to be =cre of a hope than a a scientific fact. We should note, in passing, that if the =axinun value is defined a using a probability ? (Z~_ax), then there is a very clear relation between the maxi =u= value and the rate of seismic activity. This has been described, in a =cs: obscure way, by Housner (1970). His argu=ent say be restated as fellows: let us assu=e a linear. unbounded frequency- =agnitude lav of the for: icg N = a - bM (1.10) where N is f.e cu=ulative number of events, with =agnitude > M, per unit ares, during a ""4' ** e period (per year, say). Suppose that U is the nu=ber of events / year. that can be considered negligible for risk purposes. I

16 e Then log :! =a-bT (1.11) n =ax For two different regions, with different a and b values, we have o e =abY2 2 =ax(2) log :I =ab,K=ax(1) n 11 so, h Sg " *,- o 1 o Y=ax (2) = b Y=ax (1) e (1.12) b 2 2 t It is reasonable to set b : b: 1, and then y 2 e o Y=ax( 2 ) = Y=ax(1) + (a -a) 2 1 or 2 ,7 [=ax(2) = =ax(1) + Icg (1.13) 1 '7 c I where :T is the nu=ber of events with =agnitude 0, which =ay be taken as an indication of the level of activity. In a si=ple exa=ple, if area 2 has a seis=icity of one-hundredth of area 1, then the [ x value fer =a e area 2 vill be two units s= aller than the Y for area 1. =ax The reason that Housner's (1970) argu=ent is obscure is that he tries to associate the above with a true M value, as shown in Figure =ax 1. Clearly the analysis really refers to our unbounded frequency- =ag.itude lav. In s"--a y, existing literature sc=eti=es atte= pts to postulate a relationship between seismic activity and the upper bound to earthquake ~ size, but success in establishing the nature and even the validity of this relaticnship has been essentially ncn-existent. 1.7 Pattern Reccgnition Arrreaches Recognizing the funda= ental difficulties involved in trfing to relate the size of =axi=u= possible earthquakes to the level of seis=ic 3 p-- -.. ,____.,_.._m__,,,,

j i 17 activity alone there have teen several atte= pts to include a variet; of other geophysical and geological info'r=ation. 4 Rienichenko and Dehiblad:e (197L) have ec= pared and correlated the esti=ation of K using the level of seis=ic activity, the gradient of =ax the Scuguer gravity anc=aly (suggested by Tsuboi,19h0, and Berg el al_., 196k), and the velocity of vertical =ove=ents deter =ined by geodetic and gec=crpholegical =etheds. he three estimates were ec=bined together to obtain a single esti= ate using weights of 1.0 for the seis=ic data, and 0.5 for each of the other =ethods. The results are no = ore convincing than these based on seis=ic activity alone. This paper is notable, hevever, for its extensive ecliecticn of references. Shenkova and Karnik (197L) state frequency-energy data are not reliable enough for the esti=atien of K=ax, and urge the inclusion of data en " environ = ental properties and the rate of energy accu =ulation" (i.e. Benioff graphs). However they give little indication how these t pieces of infor::ation should be tied together. In view of the interest of several Russian geophysicists in pattern recognition proble=s (see, fer exa ple, Gelfand et al., 1976), it is not 1 I surprising that atte= pts have been =tde to apply these methods to the deter =ination of M This tooic is addressed by Sune et al. (1975), =ax and an applicatica to the Carpathian region is described by Scrisev and Reysner (1976). he general ides is to look for those co=binations of observable features that accear to be indicative of the observed M =ax 1 values. The features selected include such ite=s as rates of recent l vertical =ctica, nearby velcanis=, presence of fractures and fracture intersections, seis=ic activity, gravity anc=aly etc. The data analysis follows the usual procedures.

bst of the features chosen vere found to i

var-/ strongly with M I l l t i e i a 1 a 1 . ~. - _ . ~.,.. _., _e ... ~,. -, - - -

18 The basic proble of this analysis is, however, not addressed by the autho:s. In order to deduce the ap;ropriate relationship, values of kncvn M are needed in a substantial nu=ber of regions. Since these =ax are not readily available, the authors used "esti=ates =ade by experts". "his introduces such a strongly subjective ele =ent into the analysis that it =ust be regarded as =eaningless. l.3 Other Studies ?.o recent studies should be =entioned, the first for ec=pleteness and the second because it has an interesting appreach to the problem. Caputo (1977) has proposed a cenplex =edel which purports not caly to account for the linearity of the frequency-cagnitude relation, but to predict the -av' = seis=ic =agnitude and =0:ent. The assu=ptions on which the author bases his analysis appear to be cc=pletely unreasonable, and the paper is meaningless. Smith (1976), en the other hand, has proposed using geological data 4 to cbtain a mean rate of slip for a fault :ene over the past 10's of thousands of years or longer. Then, if the frequency-nc=ent relation-ship for the area is linear, and ean be defined (see Chinnery and North, 1975; Smith's arg=ent here is less rigerous), then there =ust te an upper bc=d =c=ent that is consistent with observed slip (3 rune,1969). hith uses geological data of Esmilton (1975) to obtain these upper be"-d -- ents (which he converts back to upper bound =agnitudes). ~his approach is one of the = cst ressenable that we have seen, but

• )

t proble=s still re=ain. "here are considerable difficulties in the definition of the frequency-=0:ent relationship for a li=ited Ocne. Even if this can be esti=ated,.however, there =ust still be difficulties in the interpretation of gec10gical slip data. Slip on the San Andreas a. , _, - -. -. ~. - -

19 fault syste= has clearly been distributed over a rather vide zone on a geological ti=e scale. It is likely that individual faults could carry 4 =uch of this slip for a period of tir.e. and then it could be transferred to other neighbcring faults. To put this another way, S=ith's (1976) apprcach requires that the earthquake prccess be stationary over the pericd of the geological data on each fault considered. This is a questionable assu=ptica for the fault :ene as a whole, and =ay be invalid for individual faults within the syste=. And, of course, there appears to be no way to apply S=ith's =ethed to regions such as the Eastern US, where geological infor=ation on fault slip is available. 1.9 Discussien and Cenelusions a The basic proble: in atte=pting to dete. ine the =axi=u= possible I earthquake in a regien can be stated quite si= ply. If the earthquake record for the region has a length T years, then evidence is available i that bears on the earthquakes that have =ean return periods of up to T years, or a probability of occurrence. dew, to 1/T per year. This evidence i is not necessarily good evidence, for the largest earthquakes in the j sa=ple. i The cecurrence of large earthquakes appears to be described quite well by a Poissen distribution (Epstein and le= nit, 1966; lannitz, ,)- 1966). The probability that at least one event with an annual probability l of 1/T will occur within a pericd of t years is -t/** ?=1-e (1.1k) i St. if t = 0, the probability is 635. This suggests that in = ore than _ird cf all regicns studied there is likely to be an apparent 1 i ..ency of large events. s w -n

20 To phra:e this another vay, a 100 year record of earthquakes vill on'.y give reliable infor=atien (at the 9,0% level) for those earthquakes with c =ean return period of about LO years or less, or an annual pro-bability of.025 cr = ore. In practice, of ccurse, the length of the ea-thquake record is often considerably less than 100 years, and this applies to =cs: of the regions of the USSR studied in the quoted literature, and to California and other active :enes. Clearly, then, a 100 year record of seis=icity is only adequate for the deter =ination of ->v' u= possible earthquakes if the =ean return perieds of these earthquakes a e significantly less tran 50 years. rnis i= plies that the =axi=u= possible earthquake =ust have cecurred several ti=es during the period cf observation. In all of the literature that has been surteyed, there is no case of a specific region where a -M-un possible earthquake car be clearly defined. Even when all regions are eccsidered together in a global earthquake record, the apparent upper bound to surface wave =agnitude M s can easily be acccunted for on the basis of saturation of the =agnitude scale (Chinnery and Ncrth, 1975). Perhaps the =cs: useful contribution to this area that eculd be =ade at the present ti=e vould be the clear and uns=bigucus da=cestration of the existence of an upper bound to f earthquake si:e in just ene region, anywhere en the globe. It is necessary to add, here, that we have not atte=pted to define the ter "regicn". This is a thorny topic (see, for exa=ple, Hadley and Cevine,197h) which has been e=phasized by the ter: " tectonic province" ( l vhich appears in the NEC Rules and Regulations, Part 100, Appendix A. l l 'a'e s ' C ' rot discuss it further here, except to note that given a =ap cf epicenters for the earthquakes in a seis=ic =cne it is always possible l l to select a regica that centains ne large events. Tne validity of such I i a selection is ve-y questienable. 1 1 I u p. y . - ~. ~..

.. ~ 21 It appears, then, that existing seismic data are unable to throv any light en the questions of the existence and size of maxi =un possible earthquakes. In spite of the deep seated belief of nany seic= ologists i and earthquake engineers that upper bounds must exist, the only reasonable apprcach, given our current state of kncvledge, is to assume that these r upper bounds are at rather high levels in all areas. We r.re therefore forced into the classic =ethod of si=ple extra-polation of linear frequency-cagnitude er frequency-intensity relation-ships. This raises an additional probles which deserves discussion. In the context of the evaluation of the seis=ic risk to critical struct.es such as nuclear power plants, we veuld like to establish a vay to dete=nine the size of the earthquake that occurs with some fixed i risk probability within a given region. Following McGuire (1976) and 4 i others, we =ay usefully set this fixed probability at 10 per year. If the earthquake process is stationary over long periods of time, such an earthquake vill have a =ean return period of 10,000 years. If the process is non-stationary, this state =ent is =eaningless. However, in practice ve have very little alternative but to assu=e that the avail-able record of earthquakes is representative of the rates of occurrence of both s=all and large earthquakes in the t=nediate past and the L=ned-13:e <~.-.. The problem of stationarity is not easily set aside. Evidence fro very long ec=pilations of earthquakes in the Mediterranean area and China (the latter was discussed by lee and 3rillinger, '1978) shev disturbing changes in seismicity on ti=e-3cales of a few hundred years. The.seis=ic 1 record in New England shevs similar changes during its 300 year length j (Chinnery and 3cdgers,1973; Shakal and Tekso:,1977). Clearly this t 1

22 rair.; the pcssibility that large earthquakes =ay te associated with sc=e long ter: average level of seismicit:/ vhich is very different frc= the recent short reecrd of smaller events. It is i=portant that research into the stationarity of earthquake processes in various tectonic environ- =ents centinue. The = cst prc=ising avenues for future investigations into maxi =u= possible earthquakes vould appear to lie in three areas. First, we need =cre infer =ation on the natu e of the strain and stress fields in seismic zones. Second, ve need to improve our understanding of the ultimate strength of crustal =aterials in a vareity of tectonic settings. It see=s likely that the true upper bound is centrolled by the size of the region of accu =11ating stress, and the ability of the crustal rock to withstand that stress. Thirdly, the infor=ation frc= geological and gec=crphclogical data on long ter= fault slip, where surface faulting is visible, =ust place sc=e constraints en the largest possible earthquakes (S=ith, 1976). This approach needs further develop =ent, though the questien of stationarity =ay limit its usefulness. ,] V * * **' *' m= em -+m,9 ws

23 4. a.n.aeYS,I C: Gun-.,ven. .n.a vaa n.-.,nns s 2.1 Characterictics of Global Catalegs A logical place to seek for infor=ati0n on the existence of upper bounds to earthquake size, and the variation of these upper bounds with tectonic region, is within earthquake catalogs..There are basically two kinds of catalogs, those c0= piled for a li=ited region using data frc= a J local network, and those ec= piled for the whole world using a global network of stations. We have chosen to begin this study by analyzing the global earthquake catalog, since this seems =ost likely to contain evidence fer regional variations, if they exist. In order to be useful for this study, a gicbal catalog must have two i=portant characteristics. First, it =ust be c =plete, particularly for large earthquakes, and preferably for medium-sized events as well. f l Secend, it =ust use a clearly defined measure of earthquake tagnitude-which is unifor:ly applied to all events. As ve shall see, this turns cut to be a =uch more restrictive condition than it appears to be at 4 .4... 4 w. Several global catalogs are available. These including events 1 since the early 1900's include Gutenbeig and Richter (195k), Duda (1cj7) and Rothe (1969). Unfortunately, the global-distribution of seismic statiens was very poor until 1960, and these catalogs all suffer frc= a high degree Of ncn-hc=cseneity. With the establishnent of the Wcrld Wide Standard Seis=0 graph Netverk (WWSSN) in the early 1960's, a much = re h =0genecus data set became available. Data frc= this-netverk, i together with a variety of data frc= cther stations were analyzed by two organisations. The U.S. Coast and Geodetic Survey, and its successors the National Ocean Survey and the U.S. Geological Survey, have produced i j _, _ _. ~.. _

2h Im'.nmination of Epicenters) a fairly rapid bulletin (the FDE, or Fr ,i...r / issued on the average about 6 =en-ns a;,e .u.. - v e n t. Occurred. The International Seis=clegical Center (15 ' na; ernaen to collect all the available data, including the POE bulle:in, und tuuue a = ore co=prehensive cataleg. Typical delays in the publica:icn of the ISC catalog ranged frc= two to three years. 3cth the FOE and :50 catalog began consistent l routine bulletin producti:n at the beginning of 196h, and since then i I have =aintained the produ: tion of very unifer= catalegs. Both catalogs, since 196h, have reccrded a body vave magnitude =b i for essentially all events. This =agnitude is based on the mavimu= peak to peak a=plitude in the first fev seconds of the F-vave arrival en I short period instru=ents (cperating in a rather narrow frequency band f, f centered at about 1 ht). Surface wave =sgnitudes M, (at a period of about 20 seconds } vere re:orded very irregularly, and only in the last year or two have atte= pts been made to =easure M on a routine basis. i s I. The requirement that the catalog te ec plete forces us to focus on the l body rre =agnitude =b. yer reus ns which are utlined in the next i i ! sections, this is not desirable, but there is little that can be done a'cout it. Atte= pts to relate M to n have shown a large scatter (see, 3 b l for exa=ple, Aki, 1972). l In the sections that f6 lov ve shall cencentrat'e on the ISC catalog t for a very practical ress:n - it is available in detail en magnetic tape (the detailed POE listing is not). This facilitates a variety of ec=puter analyses of the very large a=ount of data conce: ned. i l 2.2 Earthcuake Statisti:s There are two basic representations of the statistical characteristics of an earthquake catales. One deala with the relationship between l

• "-'A$

mm_m _g_.

25 earthquake frequency and earthquake =agnitude. The other utilizes Gu=bels (1953) theory of extreces, and is concerned caly with the largest event within a given ti=e perici. Though these two approaches appear to be very different, they give very si=ilar results when applied to the sa=e data set (see, for exa=ple, Chinnery and Rodgers, 1973, and Shakal and Tekso:, 1977). Because of this, and because the frequency-=agnitude approach uses all of the data in a catales, it is to be preferred. Knopoff and Kagan (1977) have specifically shown that extrecal statistics are =uch inferior in s0=e cases. For this reason, ve shall use the frequency-=agnitude approach throughout. Gutenberg and Richter (see Richter,1958) de=onstrated that local earthquakes in California obeyed a frequency-=agnitude relation of the for=: r.i log N, = a - bM (2.1) o where N is the nu=ber of earthquakes with =agnitudes in a s=all range g centered on M, and a and b are constants. This for= of the equation is necessarily discrete (the constant a depends on the site of the =agnitude intervals in which the earthquakes are accu =ulated). In =a:r/ cases, it is = ore convenient to use -ka -- ulative for=: Icg N = a - bM (2.2) where, now, 3 is the nu=ber of events with =agnitude M and greater. C This equation =sy be regarded as being continuous, and is = ore a= enable to analysis. It is easy :: shev that if equation 2.1 is valid, then equatien 2.2 is also linear and has the sa=e slope or b-value. -Values-for the constant b typically lie close to 1.0. Unfortunately, there is no sound theoretical basis'for a linear frequency-=agnitude curte, and it =ust be regarded as e=pirical. Even

i 1 I i 26 l i i t I i leg N s j C t I linear frequency-magnitude lav I i k s s II l l M Magnitude nax 1, l. Fig 2: Ideal effect of an upper bound to ( earthquake =agnitude, using cumulative frequency-=agnitude statistics. l u_ b

27 18-2-12586 1000g u b L l_ l-1007 F o L = L A N L if I-w~ 10 ~ oz h w D L LOG N = 7.66 -0.93 M o 10 s s. W \\. T \\.\\ u. w 1.0

• \\.

C H r r \\- J r \\{ D 1 \\ 2 1 0.1 :-- \\ E k N \\ E L \\ \\ \\ i O.01 l I i I t I t 6.0 7.0 8.0 9.0 MAGNITUDE (M ) s Fig. 3: Data frem Gutenberg and F.ichter (195k). 3 J

28 1 i I i using observational data, the universality of a linear relation is not clear. }hny of the reasons for this vill be discussed in the sections that follev. In an ideal world, the presence of an upper bound to earthquake =agnitude vill reveal itself by a departure from linearity at the upper end. Figure 2 shovs an idealised representation of this non-linearity. Unfertunately, there are two other effects that can also lead to a curve similar to Figure 2. First, any measure of =agnitude based on a limited spectral band has a built-in saturation property. This is discussed in the next section. And second, seismic instruments frequently have a j limited dynamic range, and the magnificatica is often set to record =edium sized earthquakes. In this case, large earthquakes vill cause the instru=ent to go off-scale, and a =easure of magnitude is impossible. As a result, there =ay be a purely instrumental upper-bound to measureable

• i l l i l '

=agnitude for a given instru=ent.

• he effect of this on network determina-l tions of event =agnitude is discussed in later sections.

2.3 Saturation of the Ma2nitude Scale Several authors (Chinnery and North, 1975; Kanacori and Anderson, i 1975, etc) have recently pointed out that because of the shape of the spectrum of the radiation emitted by an earthquake source, any measure =ent of magnitude based on a limited spectral band of frequency =ust saturate. l i For exanple, M is usually =easured at abcut 20 seconds period. When i s i the source is large enough that fracture propagation lasts for longer than 20 seconds, the a=plitude of the 20 second radiation vill not change with increasing size, though its duration in general vill. I An example of this effect was discussed by Chinnery and North (1975). Figure 3 shows the eunulative frequency =agnitude curve for l l

29 large events listed in the classi: study of Gutenberg and Richter (195L). It appears that the listed nagnitudes are very close to present day M s values (Evernden, 1970). This diagrs= has eften been used as a basis for discussing the existence of an upper bound to earthquake =agnitude (see, for exa=ple, Ecusner,1970). It is, however, pcssible to interpret this curve in another vay. Figure L shevs a ec=pilation of recent data relating surface wave magnitude M to the seismic===ent M. The highest tvc s O points correspond to the 1960 Chile and 196h Alaska earthquakes. Ecth have teen extensively studied and see= reasonably reliable. The observa-tional data clearly indicate a saturation of the M_ scale which see=s to a begin at about M,=7.5, and be ec=plete at about M,=S.5 The solid line in Figure a is a rough for= of the M -M relation. s o At this point we can legiti=ately ask if the fall-off in Figure 3 can be whcily attributed to this saturati:n. ~4e can say this much: if the data in Figure 3 are translated into a frequency-=c=ent graph, the result is cery linear (see Figure 5}. Kana=cri and Anderson (1975) have argued that the frequency-===ent graph should be linear, with a si:pe of 0.67, if all earthquakes have the sa=e stress drop. It therefcre see=s reascnable to postulate that this is the case, and Oc conclude that' the Gutenberg and Richter result (Figure 3) can te explained as saturatic cf the M, scale. There are tv i=pertant points that arise frc= this study. First, on a global scale, there is no direct evidence for an upper bound to seis=1 ::=ent, though McGarr (1976) has argued en gec=etrical grounds that sue'r an upper bound cust exist fairly near the highest =0=ent data point :n Figure 5 i

? 30 18 2-12585 4 1 31 10

• CHINNERY AND NORTH 1975 o CHEN AND MOLNAR 1977 30 10

~ i 29 10 o E o o u 28 o. o 1 N 10 C o 3 8 o H .o 2 27 w 10 - o s l o2 26 3 10 ...l I 25

• • 8*

10 g s g .g.* 24 10 5.0 6.0 7.0 8.0 9.0 MAGNITUDE (M ) s Fig. k: Cce;111ation of c7 published esti.ates of seismic recent as a function of su: face ave magnitude M - s C

4 31 I 18 2-12587 1000"C LL ,r i-1 100 e Pr p" o r i 0 L s .m. c 0 10:- O C z L W L D L o W cr La. i I W 1.0 t c 1 P D L l =l LOG N = 17.47 -0.61 LOG M 10 10 a ,o o l I O.1 c-b i; L L 0.01 24 25 26 27 28 29 30 31 i LOG 10 (moment) I rig. 5: Frecuency - :::ent grayn :enstru :ed frc: Figures 3 and '. .I ..>me.: .m" m

t I 32 l Second, the importance of =agnitude saturation is de=enstrated. '4 hen we ce=e to exa=ine global catalogs using the 1 bz =b scale, we must expect saturation to occur at lover =adnitudes. This vill clearly wave l the proble= of trying to esti= ate regional variations in =axi=u= earth-i quakes very difficult. i I i 2.h The ISC Catalog An incre= ental frequency =agnitude plot of data in the ISC catalog for the period 1966-70 is shevn in the lefthand portion of Figure 6. Although ISC data are available for a longer period, ve have chosen to l li=it ourselves to this 5-year span in order, as we shall see, to ec= pare ,i ! the overall catalog with certain special staticus that were only operating I during this ti=e. i l The resulting plot is typical of all frequency-=3 data currently I available (e.g. Brazee and Stover,1969, Brazee,1969). There is no / l, clear linear portion to the graph, and this has-led s0=e authors to propose a non-linear relation (e.g. Shlien and Tokso=,1970; Merz and 4 Cornell, 1973; Stewart, 197h). It is therefore very difficult to deternine a unique b-value, though typical atte= pts to do this lead to high values. of up to 15 or = ore (see Figure 6). At lev : gnitudes :::y events are not reported, and the plot curves downwards. At the high end, of particular interest to us, the graph appears to steepen, and end near =b=6.5 or 6.6. No events larger than 6.6 appear in the catalog during this ti=e seriod. It see=3 reasonable to ask if these catalog characteristics are in any way the result of the stations used in the analysis. As =any as 500 or = ore statiens feed data in to the ISC, =any of then very irregularly. To examine this question, we selected a subset of 28 stations which Operated continuously throughout 1966-70, and which report regularly to 1 =

21 .s the ISO. The stations used are listed in Table 1. !!agnitudes vere recc=puted as the average of tr.cse reported by the 23 staticns, and a 1 ....<. e +.u,+ s.a.>..s...-. a,.,,..,,o .e.a. u e,. v,.,.. ..........a c 4....- <a. .m. vas superimposed. The resulting frequency-=ag'hitude graph is shewn in i t +u, 4.u+u.an. y.C 4,.. O c T.4 .., 4 ( *..h a. a..'.' d. C.'.-.3). a 7 d g. A s e ^ ^..d. wa*a se. .4. ..e I i was ".w-e.' ky a- .S.d s- +.5..a. a..*.'.- s -....a. '. ' a s e s d a.. e..* *.a.d. '"

... *. h er f s.

4 l (1977) to the 23 statien netverk. The results are shown as cpen circles. The 23 station network shows very sinilar characteristics to the 1 catalog as a whole. In particular, the general curvature of the graph i l and the fall-off at high =agnitudes are preserved. This is convenient 1 since it allevs us to study the 23 station network instead cf the whole 1 i catalog. There are reasons to suspect that biases =ay be introduced into the netverk =agnitudes by the process of averaging the reported station =agnitudes. This proble: vill be discussed in =cre detail in later i sections of this report. It suggests, however, that it =ay be vorthwhile locking at the frequency-=.. characteristics of the even s reported by 4.24 4 s *. a+.1 -.s. 7.d".a.' l j. Figure 7 shows plats of the events reported by Kevo, Finland, for , c e.;..,o. a/-.ta, ".a. s (n*.' s. b. a. ob s a..-~. e.i n...w,,,. .3

a..... a.,,

..on a. >..,,.a. . t...u.e

s.. s. d.s....-........ e 4 s 3

r .u ,...w.. yu...a u. t. 2, g a.,... ,.o-, .. a. a. n, .u..,. y,.,.,., ,...,...,a s. a"y,'.*.*.*.'.^. s*.a..da..d sn.'.'+." da. w". s '.a.. - a. --.---.'.w-a.. s *. a *..* - n. -., 5 " .".a. 7 os f This correction is best kncvn in the distance range 30 to 90 degrees, 6 l and the righthand side cf Figure ~ shows events in this dictance range. l. Si=ilar data for Fort "cresty, ::ev Guines, are shcvn in Figure 3. l i 1

k . a. -,::.: c: =.a..-..... n.- ~.:.. w. .,..,.c.., , w,. - (.L... u., 2.4...I,.... -y,-,,) wwa i A *. ' Albuquerque, N.M. -0.20 3HA 3reken Hill, Za=tia -0.26 3XO 31ue M:ns., Oregen -0.29 353 Hensberg, Ger-* y +0.20 3U! Bulavay, Rhodesis -0.C7 CA's Canberra, Australia -0.02 C*.K Chileka, Malavi -0.27 COI. College, Alaska +0.01 COP Copenhasen, Dennark +0.36 3"JR Iu.reka, Nevada -0.2h KI7 Kevo, Finland +0.C2

-i; C:ech:slovakia

+0.10 F.I Kajaani, Finland +0.1h !JU Ljubljana, Yug:slavia +0.29 M3C Mould 3ay, Canada +0.lk MCX Maxa, Ger:any +0.02

"CF N rd, Greenland

-0.1h N? - Northwest Ter-it: ries, Canada C.C0 NT.*R 5ur=ijarti, Finland +0.19 PMG ? rt Meresty, 'iev Guinea +0.10 .:n. :.=. o,... 4.a,

C1.w. a.s..4ca

-0.0*4 ?RU C:echeslevskia +0.Ch EIS Res:1ute, Canada +0.13 SJG San Juan, Puerto Rico +0.2k 20 T:nto Forest, Ari::na -0.32 TJC Ouesen, Ari: ens -0.lk U30 Uinta 3asin, Utah -0.11 . 4 x_., e, e.-,. u.. a. *,. <. a -0.,~9 a..,,I a.m x e E i I I i in i a >tw-

cac-was 1966-70 3000 ALL EVENTS 28 STATION NETWORK 3 STATION DETECTION 1000 : ..* N. ~ o SLOPE 1.49 WITHOUT STATION BIAS ~ r o WITH STATION BIAS .o o 10 0 i. i s SLOPE 1.47 N p l a o 10 : ~ -o e o } t I t i i t t i 1 1 1 1 1 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.03.54.0 4.5 5.0 5.5 6.0 6.5 7.0

• b
• b Fig. 6: Frequency magnitude data for the ISC catalog, for all listed events (left), and for a 3

selecteli network of 28 stations (right). The 28 station network is listed in Table 1. l

~ M a 5 o. 3 4 i 7 n 3 0 3_ 1 9 \\1_5 E 2 O/ u i 2 P i. c b 1 6 A m L S

o. N 0

s O i I 3 TA 5 T i 5 S d k-na h s? 0 ln i 5 i F / I* o i 5 v 0 4 e K 7 r 0 V-o r r 4 f E6 a t K6 a d 9 14 e 1 d S u 1 0 T t t ,i1 E 13 i N P g n E a S/ O m V L 5 i 2 cy E n T e L u / q L 0 e A A i2 r P G 5 O 7 L i 1 +* / g i i F

• f 0

/* i 1 5 i 0 o E~ 0 0 0 1 0 0 1 0 1 1 N

\\ 3 s 0 0 n ,7 z' \\ 1 \\ \\ \\ c 0 E 5 b \\g 9 P

n. 6 m

n O te L N t i 0 O u A S 1 G 6 I T w A e N 0 5, T 3 5 S 1 y bse 0 ro n 5 M / / t 1 ro 5 P 0 . 4 r 7 o f G- ._ 0 r 4. a M6 ta d P6 9 7 e t 9 iu 1 t 0 5 in \\1 S \\i 3 g t T E 1f \\. a \\ m N P 1 y O 0 i y E e L n 3 t V e S T 3 E i t / e _5 A r L 1 F 2 L G A O 8 0 L h n g 2 g i h F It. / t5 1 /.1I / i 0 t_1 5 0 O 0 0 1 O 0 1 O 1 I N 1

t 38 We have co= piled si=ilar plots for di of the statiens in the 23 l station network. A vide variety of behavio.' is seen. If atte= pts are made to fit the frequency g plots with a straight line, slopes are found to lie anywhere within the range 0.9 to 1 5 Figures 7 and 8 show s clearly the differences that are observed. There are two possible interpretations of these data. If the differences in b-value are real, this could indicate an i=portant regional variation in seis=icity characteristics (clearly FMG and KI7 sa=ple different portions of global seis=icity). The second alternative is that station reporting characteristics vary considerably, and the data are not good enough to define a true b-value. Perhaps the most surprising result is obtained when frequency-station g plots are =ade for the U.S. VIIA observatories. These are 3MO (Blue Mountains, Oregon), U30 (Uinta Basin, Utah), TF0 (Tonto Forest, Arizona) and WMO (Wichita Mountains, Oklahoma). The four plots are superi= posed in Figure 9 Each station has been adjusted heri:entally according to the station biases of :Icrth (1977), and s=all vertical adjustments have been =ade to i= prove coincidence, recognizing that there are s=all differences in the seis=icity sa= pled by each station. Again, only events in the distance range 30 to 90' are included. Re=arkably, these data are all consistent with a seis=icity curve that is linear, with a slope of about 0 9, up to =b=5.8, and then the curve bends downvards and approaches the vertical in the range =b=7 0 to 7.5 This relation, indicated as a solid line on Figure 9, is remarkably s1=ilar to the Gutenberg-Richter M cu.ve (Figure 3) in shape. However, s t 1 it differs dramatically fro those cbserved by nor=al stations.

Notice, I

for example, that these observatories record =any events in the range j =b" ' T =# I' 2 ' *** # *

• 8 000* *#*

1i3=*d i th* I3C C*="105*

.. -. _. ~. _ _ = _ _ = -. -. 39 llI F1352511 J 1 I O O S,", "A 'f O m' "s 100C - A o

• VELA ARRAYS C

xA ADJUSTED FOR STATICN BIAS AND SEISMICITY LEVEL 0 1966 70 1970 L i l A o

• 4 Og

'C0 r F I \\ SLCPE 0 93 y [ \\ fo a N\\ 4 r e suo \\ a \\x 0\\ ~ O V80 \\ x 0 \\ x TFO O e \\ 307 A e a WMO x \\ AA \\ b O \\ L x= g 44 N 1 e 1 t i 8 35 40 45 50 55 60 65 70 /5 m, e ( Frequen:'-station :g,el::: for four U. S. Fig. 9: / l TI A cbserysteries.

k0 2ere are a nt:=ber of i=portant differences between the VELA arrays and the average analog seismic station. Se operators of the VELA arrays were highly trained specialists, who =ade an unusual atte=pt to =easure magnitudes carefully and censistently. More i=portant, each of the arrays was equipped with a icv gain channel, which gave the arrays a =uch larger dynamic range than the average station. These points strongly suggest that the 'E.A data =sy be more reliable than regular station c repo: ts. An additional suggestien that this is the case is obtained fre the Large Aperture Seis=ic Array (LASA) in Billings, Montana. Figure 10 shows data frc= this array for a conpletely different time period (1971). The seismicity curve shown in Figure 9 is an excellent j fit to this data set (in Figure 10 this seismicity curve has been adjusted vertically for a best fit). In oner to investigate this proble= in more detail, it veuld clearly be advantageous to limit the geographical region within which i the events are located. In this case ve =ay expect a well defined i seis=icity curve, and we can test the ability of various networks to detect this curve. This is done in the' next section. 2.5 Events in the Aleutian-Kuriles Region, S e analysis of the previous section was repeated for events in the Aleutian-Kurile Island area (defined by longitudes 135 E to lho*4, and latitudes 30 -90 ). The i.portant seismicity of this area lies within I the 30 to 90 range of stations in both Europe and the U.S. Figure 11 shows the total ISC data base for this area for 1966-70. The frequency-=agnitude data do not disagree strongly with the seismicity curve shown, which is that shown in Figure 9 adjusted vertically for a l l best fit. Upon closer exa=ination, it transpires that the catalog for i l A ~ %W"W

• R aOy QQ

_ deg O.-9ap, Q. epQ q yyg _y [

al c, .,a s ota 1000 LASA BULLETIN ~ 1971 ASSUMED BIAS = - 0.25 100 *- r N 10 \\ i i i i i i 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 MAGNITUDE (mb) Fig. 10: Frequen:y-negnitude data for the Large Apem:re Seismi: Array (~ASA) in '!:ntana for the year 1971. Tne solid line is the seismicity curve shc T. in Tigre ?.

I i 12 f 1 C22-5621 1000 ALEUTI AN-K_URIL EVENTS 1966-70 ALL ISC 100 N 10 l 1 I I I I I .I I 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 MAGNITUDE (m I b Fig. 11: Frequency-nap.itude data for all events in the l Aleutian-Kuril area listed in the ISC catalog, i 1966-70. ~ l ? I I l

k3 1 4 r this area :s hen.7117 biased by the reports frem the VIIA observatories, part*cularly for lov and =oderate edents. The situation is clarified in Figr e 12, which shevs the data for a 4 twenty-five station network (this is the same network as that listed in Table 1, with the VELA sites 3MC, 20 and UE0 removed). As before, three station detecticn is required before an event is included. Scv the shape of the network erve is clearly very different frc:2 the seismicity .i curve of Figre 9 In fact, it is very difficut: to locate the seis=icity curve in any "best fit" position by vertical move =ent. On the other hand, data frc= the VELA arrays for this area shev excellent agreement with the global seismicity curve, as shown in Figure 13. notice again that the VELA arrays record =any events with magni *.udes between 6.5 and 7.0, while the 25 station network shows none (Figure 12). It is not possible to attribute this effect to the geographical locatien of the stations used, since there are 6 North American stations-I included in the 25 station netverk. 'Je can accentuate the proble: further by considering only stations in Europe. Figre lh shevs the sa=.e data fer a 10 station European netvo u, which is listed in Table 2. ~he addition of the biases of North (1977) do not change the disagree =ent in shape with the VEIA stations, but they do reduce many of the network =agnitudes. This results fr:= the generally positive bias of European stations (Table 2}. If the postulate:1 seis=icity crve (Figres 9 and 13) is real, there are clearly proble=s with the magnitudes reported by the individual stations in the netwerk. As an ex2=ple, Figure 15 shows the observations of Aleutian-Kurile events by station KEV (Keve, Finland), which was discussed earlier (Figre 7). Either the reported =agnitudes are subject i l

l t kh C22-5627 1000 ALEUTIAN -XURIL EVENTS 1966-70 TWENTY-FIVE STATION NETWORK 100 N 10 l 1 1 I l 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 l (MAGNITUDE (m ) b t i I Fig. 12: Frequency-nagnitude data for a 25 statien netverk l (the staticns listed in Table 1, with 3.MO, TF0 i and U30 cnitted). l l J

1 l L5 1 i ~ C22-562h l 1000 ALEUTI AN-KURIL EVENTS ,8 8 Le 1966-70 e A OA UBO

• a 0

0 TFO . 6 O A BMO 100 O A .O 0 N A ~ A O O O 10 p. _~ O O ^ O A.A O A O ci 1 I I I I I 06.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 ~ MAGNITUDE ( mb) i l l 1 Fig. 13: Frequency-magnitude data frca 3 VI*.A arrays for Aleutian-Karil eventa. Se solid cu ve is the same as that in Figure 9, adjusted vertically for a best fit.

I h6 i c22-5625 1000-ALEUTIANS-KURIL EVENTS 1966-70 TEN STATION EUROPEAN NET e.

• WITHOUT BIAS

~ o. o.o. o WITH BIAS 100 o. o. o = N o. o 10 i o l o o o o O. I I I l 1 n l I e= 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 l : MAGN ITUDE ( mb I . 1 l I l Fig. Ih: Frequency-cagnitude data for a 10 station European network. The stations used are listed in Table 2. - _.._. - n - ._.n

L7 . a. _2, e 3 O S. A..v, i..w. m.....Lm.. a., o, re -m u. ~. 0,, w, ar, .., (.Ortn, .,9 ) v r Ca...CJ. d n 2.. us mm ,i 3:IS Bensberg, Ger.any +0.20 COP Copenhagen, De--a-k +0.36 El Kevo, Finland +0.02

~ic Czechoslovakia

+0.10 KJ'l Kajaani, Finland + 0.1L LJU Ljubljana, hgeslavia +0.29 MOX Mexa, East Ger=any +0.02 liUR 'iur=ij arvi, Finland +0.19 FF.U C:echosl:vakia +0.0* S?J Stuttgart, Ger any +0.29

L8 C22-5626 1000_ ALEUTI AN -KURIL EVENTS KEV 1966 -70 i, I 100 N in-_<_ l l 1 1 1 I I I l-3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 M AGN ITUDE ( mb I l i Fig. 15: Frequency-=ag.itude data for events in the Aleutian-Kril area, as observed at Kevo, i Finland. Se solid curve is the same as ft these in Firres 11-lk. i l n

to 9

• * ". *. S *. a *. ' ' *.

1. . 5

.#.*.*.*#..*.e* -.."..e*. *".*.*.*.S.

v. ' a S *. S,

• ^ ******

a. ......e ^" g ..g... .. -....e 33 g 4 4... 3 a.. a 4 .w. .X., g. 4. .w. .3.... .g... 3 4. .w a. 7. ag4. ..a a.

• 4*

. /. n..',.7 g.. o.a.. ?........ 1 4... 9. .wa. -,4.. .j. .r3. A. . 4.. s. .. g a. 4A. 4. - 4

4.. a.

4.w.

. w, n..:.. n. ....a r..... ....== g. ........ s 1.... e. ('E 4..

w. 5..:-.,... 'J w &..e i f I*s..

1..4 s., t.. e 4.... g. 4. ".~.A a .. /. .ss. n r. ..'.g...#*"'*.S .".*.....#~a*'.*..*. ~..# .w. .g., .g 4. s... 4.3

• e.' "g.' *. *. 1 *. # *
• ' * * * ~ ~

^ ****-. 4 S..S.._.. 4 E. 9.f - 3.4...a..S e. . s.e s.* S..., o.n.... S .3 .v. 9

• w.

c e.....S,

1...

i. .e. s. .e. of ~.w g~. '. &n....n..',....o g &... , e.4.e 4-. .e. &. 4 e.S. '*e . 13 - ...,A.e.. g. 4... A ~

s.........e.k'..

1. .4 J.

n...

4.. A..... a.. *

..*4.4.....#~..h...

• 4.* S.

.m.

• s

=.e e w. 4. f. S S.4 %. *.., w, .a.44..a...._.t.4+,.3.. S. 4... S,._. a. 4. ....s. .a a. 9 3..g... 4.y.g, .x.*g.. 4.. eg .w. e.. . 4. e.. v... g. s...je... 3.33.g.. .. y. .4. gn.. ,e ..4,*

g. g...

a. 4 .e, ..a +w. .g4.g3 .. s.3

w. a g a..a.. s...

4.4 f. e.. .s r. s-g ...m..g .g. 4*'J.. g..., g. 4.. .g.w.4 w 43 nag 4 .3 w-k.. g.s. l g. g g,. u,, e 5 4.%., g. un S ,4.....4.. -. u.g' 4 s 4 e.g a. A n . 4.e 4 g. g. 4. sa f.. a.. og c. .%4.3 . ~.., g4.'e .....****S

• * *
• S S a'"g*

.e S... 3* s e. g.. -.... .. e. ' 4. r. ...S....... y~f.S.S, 4.. m..... 4. .w, 4..... f.. .a..# 4. 4., S.....e .w ....e.. 1 4 ..3.. a. .3 .a .w4 .x. 4 4 g..g..w....'.. 3.....-.,., ...e ......e .... s ,.3 4. o a.. ... ' g. 4 3 4...f.w.. .w. .g.4 .g4. 3.* s. 4gg3. 4.,.4..' s. .e f.. .f (*. 6*

• e '..E~. ).

~.#, e X' *"'r' ' *- e .a '*.a '

• S. S ******

E. e a.'**. ' *v' '.* '. a. *O *.".*. ? + %..

a. a. 4...e.

4..,,. S 1.,. 4 3, o,. 1 3...,....,4, ,,. 4,,, a,. u... .4 ......,.... o y.. f.... .y..... . 4 .... ;. 44 -

s.,.. a. s a~

• a.".

'..'s. -g...- g. s. c. 4 4..,. 2 s,.u ..e

4.... i. e.

f,gs... s.. ..e,-..S e., ~ *,... .er .>.....s.a.*4.., u .s. 3.. .a... g ..j g.. 3a..,,...s...1,

2.. a. g..

.w. 4....,.., ...3 4...A..e ..4.w w .4. 3., 4S .......... a.. 3 a. e.. .g C *.* a... m. ). ~s... +. s -) A.... n. ~. 4 n.. ......4.........a...n., u.e 3.a. .....~..e E.e. ....s 4 .e.,.4... "s

4.. a.,

.w...... 4 4. .4

s.,..c.4... u...

4.g. 4.

o. r.

g.... y. ...g. 4 4 '.4...,s.

• ...a..,

1. .w,a..

y 4,., .4... 4,... a. .e 4.g.... g..w. .y....

50 18-2-13521 t~d p_ g R DETECTION SATURATION m N / OR o CLIPPING c: CL z X k 9 I t-- [ l O l-I to l c STATION m G G b d s r STATION DETECTION PAR AMETERS G d 7 A E d G 50% SATURATION THRESHOLD ~ s 7 SPREAD OF SATURATION CURVE s B STATION. MAGNITUDE BIAS P ABl W OF W O mliNG R l Fig. 16: Form cf the Oetection Probability Curve f:r 1 seismic sta, tor..

5,. n.r.u.e n..e ..n. u...s, <s a ,s.u..a.a., . n. n.. n s. n. a. +, c... u..a.n u.. .as ..a a. u.., odel; thougn it =ay be One Of the =cs i=;ortant effects in deter =ining

• b.a. d.g-..u'.. a..s-a.

L,.14*..4e .-a...*.*_3.- Anplitudes are generally ceasured with a rule cr. the seis= gra=, v"..# a. h

4..*.. s a a. d. **

S a.ea.. a'..' ' s' h. - .-..a.. s *a-k..' c a

a. r.

".'a. s =a.'.' a. s '. k

• u..

g. r g a=plitude =easurable depends on the line thickness, which is typically about 1==. Cne vould expect a=plitudes of a few =illi=eters to be easily =essurable. With larger events, however, proble=s arise. M:st operators record the a=plitude, :ero to peak, c' the first sving cf the trace. When this intersects the edge of the paper, =ost operators vill ..,c**. a.n a ,+".de. n"aa^, v-a. '..".a. . - a - a. a=r ' '. *. d a. '.ec -a.s -.a. .n. than a 'ev :=, the ability o' an operator to locate the tip of the peak (w..., u. ) v..,.' da,e..d c..

• .'..e qua' *.y

^e # *

• a. - b. *. c '...- k..' 4.-.-...'.'.e, e

e. .~ ... e. .... r s,.. which is usually quite variable. And very large events, even if they do not go Off-scale, are usually dif"icult to =easure. Cn purely gac=etrical groundr, One vould expect the dyna =ic range of a=plitude reportin~ to be between 2 and 3 orders of =ag=itude (i.e. s between 2 and 3 =b ""it*)* A8 "' Sh*ll 8 h "#' it *** 8 ** h* between 1 and 2 Orders O' =2gnitude in practice, and "cc=plete" recordings of a:plitudes (the flat part of the detection probability curve) is o usually li=ited to less than 1 order c' =agnitude (screti=es =uch less). .vne. .u..,. .. a. 4. .,. o. a.u. 4.., 4..f ..z...- .u a..4 33 3 0..... s. o. a c a,...<,,.... a, e a . u... .g.r ., 1 .a.o.z,. .a . u a. s.g.z,.

u..s.,.,

.z. (the latter are seld = =cre than a fev tenths of a =2gnitude unit: see ab,e ,3 A .i s The station detection probability curve has then to be concidered

4... u......s...

.w e-....'c.s,.oca.aa a '.. "..e =. a.-. k... ~.S..a.= e a* a ' l ' ".=.. 2. a d ...... ~ - -~a-I

i ce sG in Tigure 17. Because of scattering, an event of =agnitude s vill lead to a distributica of observed =agnitudes at a network of stations. This distribution is often roughly ner=al with standard deviaticn about 0.3 a units (7cn Seggern, 1973), and its =ean (in the absence of station b bias) vill be an esti= ate of =. Ecvever, when the =agnitude of the event approaches either the detection threshold or the clipping threshold of the stations, the distributica be :=es skewed. Iffects near the detection thresh:1d have been discussed by Ringdahl (197o) and by Christoffersson et al (1975). Those staticns where scatter-ing produces a 1:v a=plitude vill not report, whereas those where a large a=plitude occurs vill report. This leads to a net pcsitive bias when the stati:n repcrts are averaged to produce a network =agnitude. Methods can be devised for including the fact that se=e static =s did net report an event (the =axi=u: likelihcod =ethed) but these =ethods are c==bersc=e, and require a detailed kncviedge of the detection probability curves. It does not appear possible to apply the= to a data set such as the ISC catalog. An equivalent bias arises at the clipping threshold of stations, although this has not been discussed in the literature. It is, of course, reversed in sign. '4 hen a large event cecurs, those stations where scattering produces a large a=plitude vill usually not report, While these statices that receive a lov a=plitude vill report. The result is a negative bias to the netv:rk =agnitudes re;0rted fer large events. This negative bias vill be quite substantial, up to 0.5 cr 1 =agnitude unit, and can adequately account for the difference between the VI'd se s=icity curve and the 13C catalog seis=icity curve. -.=ad

e7 /d o e S h. .41.e ele. g ye Cr = 3 $. 989. pao...A.%.S f l \\ 4.. My.... X.

d. a,...

E-. w w.,f 7 ..w.... 8 O.s .=n '

• a. 4 * =

1 l erv. J . w v. ..a I I I I Ehe ..a.e4 wwdC. 8 a a C+ w t. a t 4 A n. 9.'8.u s.s==. 4 *..i ga 3 G we wt .-W. e S .Eo em a eme .( e ,s. f a,.n ew v.' w. =A

e. e. a. e.4 e. E m.

r.<.%. c., wee.ra. A . e s wJ 4 3==7 J *., 4 a. p

• e o e g

t *. +.4a . e4 9

• mym e eam Je M.s a., a. e *o w.

..* 6. e s - a e

• e 4s..

4.. = 4 4 -

• J %.... J.

,- +. * . a. a

• g

$. 4...., J g w . sa.. a* .D vv

• v

<. ' G,e*J. ~J. 4 4.., 4 %.,4. 4. n.. a 3 -..a f 4.4... ...O .w J e, e,. J

3..,3

,,.=.==.*2* q .w .e ,.3 4**+,a ,w. ..\\ m...., ,s

=- 1 I Sk 'a'e can illustrate our argunent by using da*.a frc= a single station. Figure 13 shows the data for EUR (Eureka, Nevada). The left hand portion of this figure shows a ecnventional interpretation of the reporting characteristics of this station. An arbitrary straight line is fitted 4 to the data, and detection and clipping thresholds (indicated by arrevs) are determined at =b=h.5 and 6.3 respectively. In the right hand portion of the figure, the VELA seianicity curve is used (EUR is quite close to the observatory 130). In this interpretation the station fails to report =any events for = greater than 5 5. The thresholds are new h.3 i and 6.1, and "ec=plete" reporting is linited to the range k.7 to 5 5 A 1 similar interpretation for staton KEY using Figure 15 suggest that this station carries cut "cc=plete" reporting over an even smaller range, i perhaps as little as 0.3 =b units (fr = 5.2 to 5 5). A different representation of the sa=e phenc=enon for station EUR is shown in Figure 19 Here, for each interval of 0.1 =b units of U30 1 reported =agnitudes, we have averaged the difference in reported =agnitude between EUR and U30 for events in the ISC catalog during the period 1966-70. The theoretical interpretatic'n of such a data set has been l l discussed in detail by Chinnery and Lacoss (1976). If the detection probability curve for EUR vere heri: ental (Figure 16) then this plot c should be horicental teo. The presence of a detecticn threshold shevs. as pronounced positive biases as lev =agnitudes. There is a hint of a flac portion of t5' curve in the vicinity of 5.0-5.5, and then the data centinue tecc=ing =cre negative. This must be interpreted as being due to a clipping threshold. In general ter:s, Figure 19 is entirely consis-tent with the right hand preferred interpretation of Figure 13. l 1 i

c22-ss87 EUR 1966-70 T 1000 ; l g* 8.%s /*~- / 100 : g

• \\

N ~ * ~* ~ .\\ y \\ j \\' l 10 : \\ v \\ \\ \\ a\\ ~ \\ ~ \\ a ~ \\ (* y. .1 i 1 __ _ 1__1 t 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.d STATION m S ATION m b b Fig. 18: 'No interpretations of the reIorting frequency of station DIR (hareka, Nevsia). The right hami interpretation is preferred.

I EUR - U BO i 1966-70 x. l e g i e s.e q e i e e e n -s r ; t,l e l I t e i i e, e e i

i. o ee, e

l i.J e l ;! ee e e j i s e e -i ,/ t \\ s e I e I I I I I I I

5. 5 4.0 4.5 5.0 5.5 6.0 1

MAGNITUDE (UBO) I Each point is the average difference tetveen the station magnitudes of IUR snd U30 for al' events listed in the ISC

staleg, plotted as a ^anctica of the U30 e.agnitude.

j t-e ~....- -. -....-.... _ _. a _ _w, - 9;. g li) $ n l M_3 y. . T..pQ .. ~

^n.MQi.WMACGGyl

..........,.. w,;,_ A A .M I

'Sf. A.T-$.l~,)~b, ?.+ 5?.l. hl.75%*WA/;240'.W kf,

'22Th'ntf i* 4.'. ?.,h.h.. ji'" "' j' . - f 't f.{,.'.:.f$.. g.. ,g %, 1., a. j.,_,.g ,s* s 'h

• • ) ?; h,

4 4 - G y QQ.t a.3,.u -?, g4 .p . ~. ' a -.gggr d-gu g,A" ,=,' *,.,, g *-

• 4 bp

, g g m?'bgu n e' '. J~ ",*C g _,,'~. 4h: iM.,?:..'p Qf3C..%+'. 9 .r.. -k ', y m p r .w p~ ~ e M D ...-.u r..:.'t K.n!.*4f.4 &si . -. h.' "eM.Q}***

a . =w )e o.- 4..,,,,S...4... a .a ".w...,._S,.,.9.. 4e3.5 4w.4 g h.,......,7. 44.. .....J..4...... a..y 4 4...... w.s. 4 3....., .a c.. 3. ,, 4.

4., o.r g. g'

. 3. g. 4 3 e3 g .e.. y... g. ...ws .g7. 4 .w3. ...yy ...6 ga. -3.. 4... 4,3

• *......y...g

,o,gw. 4. 3 w... 4

4...,..e

~. h .1 1... 3.. w..,aS.2 3.o. ... e....... . 33 a4.. e.14 Sw. ...'.*3... a .w4 3 4 4 4...S..

• S.w..

4. 4,. ..e....*a.. ....S n.... .s .a ..s 4.c g a-e s.1..c 3. 4./. e ae.3 4 4.y w g g.. 4 g.4 3.o. r 2 4.o.rg..... g 4.. 3 .w . 4 .e... 3 ( 3.4 4 .r. } g.,,3.9 .g, S.4 ... 1.,w...g. 4.S ",.,.:....a." wy & .. S.4 .w4 4 S

a. a.r.s.....

.e g. s 3.g.4..g,

4.. s.. 4 4 s.e..... g...g.4 3*. 4.. o. 4.e..wg g.g..

6 .wr. y.. 4S.4 a. "v ve* r k.. o 4 a... 4 . w. 4.--. -4. A. #. L 4 4 E ...s E...a .o .j r f..... p. .o.. .oe. ..a,..4.* g 53 0.g.1 s . -....,..g,. c d.4 4 3-g.. 4,. g 1 e * "..

• 3 1". S *# ' 1'" ' *-

a.*. d 9 s. b.. w S. S &.* *. ".*" '.."4,. e d. S S U*, S *.1.".*. #.1' .d a*.3..t, *..".*. "v.'u'.A&*.*

• a* *** C"..N " e.' d. 3 ' *.

k. A ...... g. 3,.... v.# 4 s ,,.3 4........... e w... a.. .e. s. .g, > g... A 4....... 4 .yy .v ..s. .a...... 6.4. a7 ..g 4.g. 4. a. 4 3 e.45 4.4. y o,.3. 3 k.. egg' S.a./a. .j... g.7. 2......s. 4 4 n3 g4. .X,a 1,, T4.3 .M. v a.. yg .g. g e.-. g w..gi.a y. 4.e .4 4..f e.i .'e .w a a.. -..,....g c% .w A. e = '. a. a.a. S " k. d." a. *. # ^ a.. -... *.. ~.'k.. a. 3.'.- S., s.'. - "... (?4."..a. 9 ) d. a a.sn a. .3 s . X. 7.*.. es. . w. 'n . a 3 4 g 4. 4... r- .., S S S.. .wS.. .s -....e .3...

• 4 4 ga

....,gg 3 .... 4 n. ..r.. .s a,1....

4.,.

.w. oi., .4 ..e ,1....S. a _ a.1, w.Su e 2.,...-,..3 -.... a 0 n ...v 3 \\'.i Q-C '.;/ +0 ,.. S..n.. En. ...... 7.3

4.. S w. 4.,.,

r.w a... s, 1m.a.."3 0.. w Aa.30 4.. .w. 9.,4 3h.4., ..v.. .......... a. .of., _3. 4., 4,,. 45 -3 3...., J .I,Y._ s..,.a \\.)..) 1 3 e a e.,~. . 9--L':, .w. f.e., ?e-w .y. e. w., .w -.e sa. w ..... 4 3 a.

s. >..

4

4....

..t. .c.. 4.. w., S.,... .w. 4 a ,.3. ~... 3.. a. ...,g.i.,, 4., ... 4 .). k..

k..

4.... g *.i 4.3 4.. 3. '. C.-.e a u .s

4. a

.. e.*.* 4. P J. o.. g A.. a ..w. a

53 -~ :-. c- .u- = 0 1000 SOUTH AMERICAN EVENTS 2 1966-70 obga UBO o .o o TFO O ^ A BMO 100 o . A e N ~ o O l l A O O 10 o e l 2 o 3 O o A l O o ~ l 1 e i i e e.g.x 3.5 4.0 4.5 5.0 5.5 6.0 6.5 6.0 MAGNITUDE (m. ) g Fig. 23: Frequency-nag.itude data f:r Scuth 1rerican e' rents obserted a: 3 '/ D a rs:ts. 'he 3:1i1 : r t= is the 3* a 23 that in Firre .0, -djus ed retti:111;r for a test fit. l l

59 C22-5622 9 M ~ 8.6 s 8 M ~72 s\\ [ 7 m b g \\ m ~73 b os m ~ 5.8 b E \\ o6 g <t 2 y s 5 li 4 I I I I I I I 22 23 24 25 2,6 27 28 29 30 31 l LOG (moment) Fig. 21: An e=pirical =b-mement relationship consistent with the VILA seismicity curve (Figure 9). The M -mement relation-s ship frc Chinnery ar.d ::crth (1975) is shown for comparisen. f 1 f r 4

61 Ru aENCES

• References =arked cy an asterisk are included for co=pleteness, but vere not used during this study.

Many have not been translated into English. Aki, K., Scaling law of earthquake source tl=e-function, Cec;hys. J., 31, 3-26, 1972.

• Anan'in, I.

V., Assess =ent of the seis=ic activity and the =axi=u= possible energy of earthquakes in individual seis=ogenic zones in the Caucasus regica, in Seis=ogenic Structures and Seismic Dislocations, VNII Geofizika, Moscow, 1973 Archambeau, C., Estimation of non-hydrostatic stress in the earth by seis=ic =ethods: lithospheric stress levels along the Pacific and Nazca plate subduction enes, =anuscript, in press, 1978. Eath, M., Earthquake energy and =agnitude, in Physics and Chemistre of the Earth, Volume 7, Perga=on Press, 1966. Berg, J. W., Gaskell, R., and Rinehart, V., Earthquake energy release and isostasy, Bull. Seis=. Soc. A=., }h,, 777-78k,196h.

• 3onilla, M.

G., and Suchanan, J. M., Interi= report on vorldwide historie surface faults, Open file report, NCER, U. S. Geological Survey, 1970. Scrisov, 3. A., and Reysner, G. I., Seis=o-tectonic prognosis of the J =axi=u= =agnitude of earthquakes in the Carpathian Region, Izvestia, Earth Physics, no. 5, 21-31, 1976.

• 3orisov, 3.

A., Reysner, G. I. and Sholpo, V. N., On the preparatica and use of geological-geophysical fata for the identification of :enes with different Mnax values in the outer zone of the Alpine folded region, Sy=p. on Search for Earthquake Predictors ( Abstracts ), MGGGS, MASFZN, Tashkent, 1974. Brazee, R. J., Further reporting on the distribution of earthquakes with respect to =agnitude =b, Earthquake Notes, LO, h9-31, 1969 Erazee, R. J., and Stover, C. V., The distribution of earthquakes with respect to =agnitude =b, Bull. Seims. Soc. A=.,12,1015-1017,1969 I Brune, J. N., Seis=ic =ccent, seiz=icity, and rate of slip alcng major fault zones, J. Geophys. Res., 13,777-78h,1963.

• 3une, V.

I., Kirillova, I. V., Anan'in, I. V., Vvedenskaya, N. A., Reysner, G. I., and Sholpo, V. N., Atte=pt to esti= ate the =aximu= seis=ic risk: example of the Caucasus, Proble=s of Eng. Seism. lk, Science (Nauka) Press, Moscow, 1971. no.

62 0 33 -, e 4-c. w. =3x. 4. e-1 .m., 3 a..:,o,y _.>.,..., - ov ..,a.4 3 vS ..w w. e..., v s, y .w..... C a"- a s"m.= .- a. 3 4... E... d a s '.a "4 - r v4.*.k.. seis-.4e .a .w-,.af..o-4. .-s. c n >.> +. 4. s n s, "3 4 4 nv. a s *', a,,.4.4.y, .4n :ny_s.4. a 4. . e. .c. 4. c .4., w ,-,i .s.4 s..., nev, .fr.,

• n,.e, "d.

m. s.C,

4. C.w.,

_2-.*swV, 2 A., G4**s, 7. G., "..e"-.a..*, s 6.. s o. a ne a *.., o g1c s..a.P w G..T., 3 a v....>..,., ?. ?.., v...w-a c. meV,.1,. c -5 w.- ....w s . v y woe +...... .w. ...+ w, uaV.. a 4.... a....a .5..e.e... *c "yara"*ta.. s c# .... + =w a....... 3.. ...g .~ o 7_ L, 1 y s.L. .3-v.., C..-.. aca. c 4.. f. qq 4a. a g v-.., A,, a.4 4, "4. g., p.ygne,

..- '/.

?., - - e y 4.,', T..

m..,

m_ 4.e.,y 2

m...w

..o F. aca v.... > -.I, ?. ?.., v.... w e a. C e...n w s. 4

a. 4. g.w. - n y 4 ~..-

v.., -w f w3 a-.4.".da. c# e a.*.. s".a'4a. o, ..~.=..=.4.a, Ta--k

ky*4. s,.....'O, 31 k1, '9's5.

s...

Caputo, M., A mechanical =cdel for the statistics of ear *hquakes, =agnitude, =csent and fault distribution, 3u11. Seis=. Sec. Am., _6~, Sk9-861, .fil. Chen, W.-?., and Molnar, P., Seis=ic nc=ents cf =ajcr earthquakes and Central Asia, J. Gecchys. F.es., _82, the average rate cf s7 da e9ty-c9,e, .9is. n s Chinnery, M. A., Theoretical fault nedels, Pubs. Den. Obs. Ottava, ,eg-o,, c2.-cc3, ty-i. Chinnery, M. A., and lacess, C. T., Magnitude differences between station pairs, in Seismic Discriminatien, Semi-Annual Technical S - a y, Linecin Laboratcry, M.I.T., 30 June 1976. Ck...; .a., V.. a., au. "..-..k..,

7...

"s., ".w..e '. ". a uaa / ^f va / 'ar3. e a*. *~b-d -s quakes, Scie la, 190, 119~.1193, 1975 r_ o.. w,,..

u. s...<-+4,s 4n cu+.w...rn

...... - e C'. 4.. e., v... A.,.. a. e..,s a -.. s,

n..

a., .....,,_x 3. s New England, Earthauake "c*es, hi,39-103,1973 ae,^ss, "... ~.., a d. - C k..d...m- +..- "s, V... a'., S *.a.. ' s.i-al cw._. 4..,.. r..e.. s s.-., r. a., ..o e,>,_...4.. m4 c.4,4.a+.4n, 4 ., d,.., .ec. -,,g.4.

a. s. o.....4. a. 4. -.,

-o m.., . - -. ~. - ..Semi-Annual Technical Sun =ary, Linecin laborato y, M.I.T., 31 December ,c e. .s s Cp4*-3** Ccc

• A **
• 4 4

S 11 A *, T 6,*.. 4 5-3.ie 4 c .4-3'f* mea 7 3*g9 "*

• F..e19 P
  • • • *J ww..

?c; A6 a

_2.3.,.. -.cC,,9cc. c.,....,, v. a., a. a. v..., =... .s., e.4. 4.

4. ex

3.., -.4 e or ueo.0.,

-+ -. s o. ~.4. c...a+.,4., C.u. e n, J. a.e v :..,

4. n. a_,.e. _- +

1 -

..>...4.,, v.,.,.. :.,

s .C,*,.*.

a.,4.,

a c " n ^ o

• v-..

c a d e.4 -.i.s> a a.'~,o.o 4-u....,', C. a*.,.'..d "..a.-, "... a., v 4- _c-L:, 3- .e..

n. 4..r., n_e -r, - 3_,_,,.,.. -..g,.:

c ve. .v y. u. Cosentino, P., Ficarra, V., and lucio, L., Truncated exponential ..~s .. s. s ..,,,2..g u...<. 4.,. e.,..... s, e.... a. a'. '. s + 4. c s, *"~' '.. c.. -,.,,,.. g3.. 4.. 4. y. e.4.-. e,c.m., s',.o'1:.340, .Srr a f 12 .s

el Ti: u.E' ICES

• Peferences tarked by an asterisk are included for ec=pleteness, but were not used during this study.

Many have not been translated into English.

Aki, K., Scaling lav of earthquake scurce time-functien, Gecchys. J.,

31, 3-26, 1972. ~

• Anan'in, I. V., Assess =ent of the seis=ic activity and the =axi=u:

possible energy of earthquakes in individual seis=ogenic :ones -2 the Caucasus region, in Seis=cgenic S**"a+" es and Seis=ic Disiccations, VNI! Geofizika, Moscow, 1973. Archa= beau, C., Esti=ation of non-hydrostatic stress in the earth by seis=ic methods: lithospheric stress levels along the Pacific ui Nazca plate subduction zones, -a-uscript, in press, 1978. 3ath, M., Earthquake energy and =agnitude, in Physics and Che=istry of the Earth, Volu=e 7, Pergn=cn Press, 1966. Berg, J. W., Gaskell, R., and Rinehart, V., Earthquake energy release and isostasy, '" <eis=. Soc. A=., Sh, 777-78h, 196h..

• 3cn111a, M.

G., and 3uchanan, J. M., Interi= report en verldwide historic surface faults, Open file report, NCER, U. S. Geological Survey, 1970. Sorisov, 3. A., and Reysner, G. I., Seis=c-tectonic prognosis of the =axi=u= =agnitude of earthquakes in the Carpathian Region, I:vestia, Earth Physics, no. 5, 21-31, 1976.

• 3crisov, 3.

A., Reysner, G. I. and Sholpo, V. N., Cn the preparation and use of geological-geophysteal data for the identification of enes with different M 1x values in the outer zone cf the Alpine folded region, Sy=p on Search for Earthquake Predictors (Abstracts), MGGGS, MAS?Z't, Cashkent, 197h. 3ra:ee, R. J., Further reporting on the distribution of earthquakes with respect to =agnitude =b,1Earthcuake Netes, ho, k9-51,1969 Bra:ee, R. J., and Stover, C. W., The distributien of earthquakes with respect to =agnitude =b, Bull. Seis=. Sce. A=., jg, 1015-1017, 1969 3 rune, J. N., Seis=ic =c=ent, seis=icity, and rate of slip along =ajor fault zones, J. Gecchys. Res., 13, 777-78k, 1963. +3une, V. I., Kirillova, I. V., Anan'in, I. V., Yvedenskaya, N. A., Reysner, G. I., and Sholpo, V. N., Atte=pt to esti= ate the =axi=u: seis=ic risk: exa=ple of the Caucasus, Proble=s of Eng. Seis=. Ih, Science (Nauka) Press, Moscow, 1971. no. g

-t. ('O S e.".*..*.".w", U., a.*.d LU.' 4..*.,, D., a' e' a..'..." a.' #.".a.'. ' 0.". v'.

• b... %. sCa..". s" "%..4'.'*..

w w + %... u... ,n. u... s.4., c.e a a,4 -.. .s,.4,. ~

e..,. 4, n

.4,. _2.,. 4 m, a, m .a. .v yv .. ~...... Geo*is. (?:ce), 29, 1-2, 1976. Bav. 4..S, e. .nd Q,..n u*. .. 3. w o '. .s 4 e ' - A'.

1. . r 1.

.A a '. 4. e 4 .. - 6 y. L. a* v w' u a r me.4. 4,4+,

g. a....,

.M_ a,, ' s 1 * *"*i, - 0... au.. .v. B.wn p/g, if., g.e.a w.,. p a. %... '., ."4.

v..,

.V. .e.w. c. - y 4. .,,y. g 4 b.1,. ,,, gw. v.. .m.. n. .y n,w,a*4 S c1. v... t.r.,.o n,,. e a a.. a. 4 a .43. c.., %... s-. 4..o., ro.. k.. ow... i., S,, J,.

1.,"),

4 ..a. w s ..._.s "li-f", noto. w D .s

• d.e, S.V., wea.l.1a= a..iS.4C.. v..o w.1.o..

O

4. 3. k.. Piv.n.o. Paw *4.'4.4.3 h e.1.,

D' ...... e,y .. w. - .L a 9 $,.0,.i c,0's,.

m. n..,,.. Dw..

4. u.., -o, ea v

• 0.,h.4 D.1 aw.a,.,..

x., b.,. 4.e. i, a .,.4.4.... <..f S a..i. ay 4.. ca,.+ h,s,,,1* eS 4n 4 a. .4. t %... ".'e r~ i *. o =. ~ v~ * "a *. o. e~.* S. c.~.^. *

• s "v ' w~ *.". "'."s, * ~.
• . '.u^.~, O *

.= *..s~. w"~ g Canger, " Fan", TaShkent, 1971.

pS+,.4 a.4 c_.4.

, c.., n. o 4.,. c c,_....,.e.o.m.e o, 3 ,e.. r+.w... u. .w. quakes, Natu e, 211, 95L-956, 1966.

• ESteta, L., Seismicity prediction:

a 3ayeSian approach, Proc. '.t h W.C.E.E., l_, Santiago, Chile, 1969, i

. e. 3,

ewe,. 4, ,.,....aK a 4. 4.

4.,

4 4.. w.,.,. 4. 4a.CnS, 4.n g g.s..,4.u, .s. S ' s*n 4 s..S.3. a. a. ...s e.

g... 1 a. g.

D. og... D. 9.3.

• e,

V.. *. 's'. O .a..,

• C, 's O.

.s. a.. .. s. i r,/e.,a.n, J. p S +,.,a4./ c..4 o. 3.n.o..a.i S. 4.e.. 4., 4..,. a r.a.. 1a+oe, ,.oc1e.S, a f. ~. 4. m. c,,,. * ,3 .m. y.,, - '..", oy.v. 5 2

• +-n

.o.,.,1 9.. c~ ru.., .s w VayS *.4,f, ". "4., ...v...,*, r.,.w..e n.

4. %.. e,,, e.c

+ %.. . %..,.,.,,.. O u#' .a K w.., n. .yy. ..w.. w g 1 o..p. . e *4 * ' ue S '.h..*. * ** *'.**a~ ' '. ' '.**.*'r'.**.".".*s^.#.' 5 e

• S '** b d.e S
• * * ~

.w y.. w '*4 / 4.., c.s., S S.,. i, 4..., 4. .%., n 4. g .s.. %.. _ r.a .we. e o*..-* o .... s vla., 4t.g e,wq., $.a., g.,.,, s.. McSkov, 19c5 I i (t' a..i.#g 4, ?. I (F,a, k e. a., .C. a. 4

7..e..*.2 4

a.".i..m W...f., i'. f.o...w. l a ..,6 .4. n., e on 14 a . o. n. ae..4 4.O p o w. v.,.. o ' a. c, i a...

• e.* S 4.=.

va7 4..*w . e.a, w .c .. yy..e. a -... s. 4 n s.. d J. O. k...f g. rg. k. c l o.. ?. g' 7, 2.0.*a .2,,0,, 7.gi.. . i GO, wl.,./, tl., n* . X ,w.. a. w..a.,.. . p.,,,. +%...,. 4

• • .. w

.i.. e ~..s.a. .') w.,.n,. a w ....g. y.,. ?. /. S. 4..,.r.a .k.. O. k. i S., 3,.,,. '.9 0.' L, 7 e 'C.. 4 u. fC I g,.-b ol, .r.. .. g. 4.

4. o g 4...,.

4.,,j...:.,.u. a., 3 c, c o, a. ,.,e r_y.,.. a., a.,.. s 4.3 4,. i v., ......a. y n..u.,,e. %,. , d a.n 4

3. 4 Wo....,

T. r. 4.

4.., 4.....>

. %.. %.. s. o.. : y r~ 4. 3, v. s e. -~

t. v... -.. n,

0.,. 4, o. g. r. r.' 4.f 4.,..s ..o..,a, = c, :e.' tw ..a. .e.,n.a 7,,f, d. r., .o., 4 n.s. 4. n.,., u. 7., + 0. 4. - -.. e,,..,.,, 4,, a. .*. %.,. r,,a..,.,.., s a-. y v. w... a.g.... v...,,, ',,.o r,i,,. y^ ~a *' t I.- e.,., p>. c. 'se C.w e. , 1 0.....'., $.M w 7.4.74 a im 2 s. -.. Mu.g,.4.i D., 74 a1

e. s e.. * 't

".".a' s* *a*.# ~ "e"..'.** ",iak.' ", N.j ."...'*.a. .D,,,k* e

• o

'w"","*. .A.. a .v. g

t..* a..,

n*,y.4. 4 4 g 9. c.a

...a.

2.. ig, y.

O. q*o o'.s, C' 4, 1393 ., t y ... s.o, s L I L_ j _y-g 1

6b

• Hor = ann, R.

3., 0;2*e of the art for assessing earthquake hazards in the United St1tes, Rept. No. 3, U. S..Ar=y Eng. Waterway Exp. Station, Vicksburg, Miss., Misc. Paper S-73-1, 197h. Housner, G. *4., 2esign spectru=, in Earthquake Engineering, ed. R. L. Wiegel, Frentice-Hall, Inc., 1970.

• Kallaur, T.

I., Seis=ie activity and the energy of =axi=u= earthquakes in so=e regions of Tark=enia, in Study of Seis=ic Danger, " Fan", Tashkent, 1971. Kana=cri, H., and Anderson, D. L., Theoretical basis for some e=pirical relations in seis= ology, Bull. Seis=. Sec. A=., $1,1073-1095,1975 Kana=cri, H., and Cipar, J. J., Focal process of the great Chilean Earthquake, May 22, 1960, Phys. Earth Plan. Int., 9, 128-136, 1974. Kans=ori, H., and Jennings, P. C., Oeter=ination of local =agnitude ML frc= st:Ong-=ction accelerograms, Bull. Seis=. Soc. A=.,{8,,h71hS5, 1978. Karnick, V., and Hubnerova, "., The probability of occurrence of largest earthquakes in the European area, Pure and Appl. Geophys.,10, 61-73, 1968. Knopoff, L., and Kagan, Y., Analysis of the theory of extremes as applied to earthquake proble=s., J. Geephys. Res., 32., 56kT-5657,1977 Kogan, L. A., and Shakirova, T. D., Assess =ent of Keax by =eans of Gumbel's first =axi=u= distribution, in Proble=s in the Assess =ent of the Seis=ic Danger, "Nauka", Mosecv,197h. La=bert, D. G., Tolstoy, A. I., and Becker, E. S., Vela Network Evaluation and Autc=atic Processing Research, Technical Report No. 7, Texas Instru=ents Inc., 9 Dece=ber 197h. Lee, W. N. K., and Brillinger, D. R., A preliminary analysis of the Chinese earthquake history, paper presented at the U.S. Geological Survey Conference en Seismic Gaps, Boston, May 1978. Lc= nit, C., Statistical predicticn of earthquakes, Rev. Geophys., h, 377-39h, 1966.

McGarr, A., Upper li=it to earthquake size, Nature, 262, 378-379, 1976.

McGuire, R. K., Methodology fcr inccrpcrating pars =eter uncertainties into seis=ic hazard analysis for icv risk design intensities, presented at Int. Sy=p. on Earthq. Struct. Eng., St. Louis, August In,C. yl Merz, H. A., and Connell, C. A., Seis=ic risk analysis based en a quadratic magnitude-frequency lav, Bull. Seis= Soc. A=., {3, 1999-20Co, 1973

k ..o) 1 i Milne,'4. G., and Davenport, A. 3., Earthque.ke pec'estility, Puts. Er r. Obs. Cttava, Seis=. Series, 1963-L, 19;p., 19(3. Netnhofer, ".., "en-linear energy frequency :urves in c*a:ictics of __72, 76-53, 1969 earthquakes, Fageoch, "ev= ark, N. M., and Rosenblueth, E., Funda entals of Ear-houske Ercineering, n.e,. 4.

a. 4,- r,

... F. .9 7.,. Nordquist, J. M., Theory of largest values applied to earthquake =a5nitudes, Trans. Am. Geo;hys. Unicn, 2_6., 29-31, 19k5. North, R. G., Station Magnitude 31as. Its Oetermination, Causes and E:':Sn:. j Lincoln Laboratory, M.l.2., Technical ::::e ' 977-2.,1977. Otsuka, M., Cut-off of seismic energy, J. Phys. Earth, 21,119-123,1973 Papa:achos,3.C.,Dependenceoftheseisticparameterbont$emagnitude range, Pageoph, 112, 1059-1C65, 1971 P'ei-shan, C., and Fang-lui, L., An applica icn of statistical theory c: extreme values to moderate and long-interval earthquake predicti:n, _16, 6-25, 1973 (Plenu: Pub'ishing Corp. Acta Georhys. Sinica, translation, 1975).

• ?erkins, D., The search for maxi =u= =agnitude, "0AA Eartha. Inf. Bull.,

18-23, July 1972. l

• ?utrushevskiy, 3.

A., Cn the relationship between the earthquake 0f =aximu= intensity and the geological state, Bull. Council Seiss., no. 3, 1960. Richter, C. F., Elementary Seis Olczy,*4. ". Freeman and Cc pany, 1938. o-.r.4. g;,.e,

m..,

e' -.a+ 4 s*.'. s o#. ul. 4._- o * =. ~=~.~.2'.o". '. k.a. x. a a ~. ~. ' ' s c..-..- *. m=.a. u d n s trobability of earthquake cecurrence, Teet nophysics, 2_6., 1-21, 1975 Ringdal, F., Maxi===-likelihood esti=ation of seismi =agnitude, 3i01. _66,'769-3C2, 1976. Seise. Sec. A=., Fi:nichenko, Y. V., ? ssib:11 ties of calculatine naximun earthquakes. Trans. (Trudvl Inst. Phys. Earth. Acad. S ci. l'33?, _2_5, - 192,- 19di. .e. 2.,.. 4., >.., o, y

e..,

2.0. 2

4.,..., 5. 4 -

.y.........s........>.....-- ,f + m .~.. ......... e s s. .....e,. a. ._... u. p, ~,.r..,...a g g.< o. .< n g..> ~. 4.,. m., y._ u.n. v:., a.. "..a..,. na m - ~.o., m.. .a , y m.a. ,,,_?,-;7., .c Ri:nichenko, Y. V., Oeternination of the energy flux' of etrthquake f;:i o...w. Das4

o.,S.4. i s~. 4.,

S.. 4,.4.., D. u, v.. _

4..

• ,,4

41L, K e _e,

_ c,. e.. a -_.e..-. w 30.' 7.2',.' 96h w. v w p* ~ W ,g*P t?9l" ,-n.'. ,%*,"\\.'

• i" -

I

66 I

• Ri:nichenko, T. V., Seis=ic activity a:.d the energy of' the largest earthquakes, in Problems of the Gecchysics of Soviet Central Asia Ka:akhstan, Science (Nauka J Press, :'esecv,1967.
• Ri:nichenko, Y.

V., "he generalised lav cf earthquake ccourrence, 3c11. di C-ecfis. Theer. ed. A; lic.,12, no. L8,1970.

• w 4. _4., w.......e.,,,.e.

y., m.w..-... - a. s '.,e s.- '.M.e ea-*. ' -" a b.. s, T.a.h ar.d Universe (~e=1ya i Veslennaya), no. 5, 1971.

• Ri:nichenko, Y. V., Dete =ination of seismic danger, i: Problems in the 0.uantitative Assess =ent of Seis=ie Danger, "Nauka", Mosecv, 1974 O

Ri:nichenke, Y. V., and 3agdesarova, A. M., "'he strongest possible earth-quakes of Japan, I:vestia, Za-th Physics, no.11,1L-32,1975 "Ri nichenko, Y. Z., Dru=ya, A. V., Stepanenko, 3. Y., Crustal seis=ic activity and the maxi =u: possible earthquakes in the Carpathian-3alkan Region, in Regi:nal Studies n Seiscie Regi e, Shtinitsa, Fishinev, 197h. Ri:nichenko, Y. V., and D hiblad:e, I. A., Dete.~1;ation of the =axi== possible earthsuakes on the hasis of the ec=prehansion data of the Caucasus Region, I:vestia, Earth Physics, no. 5, 6L-35, 197h. Ri:nichenko, Y. V., and Zakharova, A. I., Generalized law of earthquake cecurrence, I:vestia, Ia --h Physics, 20. 3, 29-33, 1971. c .w._ _.... ocn 394g g.< C0 ""~%_'.*ca 'oa., n. w

2..,.. e. i s-.4. 4.. y e...

...w .-re ..e,v. ...~ 1969 Shakal, A. F., and Teksc:, M. N., Earthquake hazard in 'iew England, .e - 4...., .c, c, ,s,._,i3,.-~~i. y ~ _... - I Shebalin, N. V., The -M = =ar-itude and =axi=u= scale intensity of i an ear-hquake, I:vestia, Earth ?hysics, no. 6, 12-20, 1970. Shebalin, N. V., Isti=ation of the site and position of the focus of the Tashkent earthquake fr:= =acreseismic and instru= ental da a, in "~he Tashkent Earthcuake cf 1Me, Uzbek 3 ranch, Acad. Sci. v. e.e. -,..s , _9s,. -... o 4 S.hebalin, N. *4., Assess =ent of the =ani== seis=ic danger in the Cri=ea-Ts:s:sk region, in Seismicitv. Seis=ic Eanzer in the Crimes a.d Seis= stability of Stru: ".*es, "naukevaya Ju=2a", Kiev, 1972. 4 -..,v.,., y., m.......k a "_ * * - o.# c". ur e ca c'. lar,as* x. e....,-..,.,.,.4 t.... .s e earthsauakes in the Zur t.ean area - Part II, Pu e and A 01. Geophys., er. ao ,c- ,9en. n, 2 s 4-2,, v Shenk:va, 2., ani Karnik, V., C:=parisen cf =ethods of determining the largest possible earthquakes, I:vestia, ? "-%

• ysics, no. 11,

. 3 ,1 d.e, of6*. / I I l D" ~. -.......... ~. .. ~.

.o7 -...c.. 4.,., ..v.,.,. r..,...-.....,e-.4.,,2 a. 3.

4. a-. 4,. -

• v w

. v... y . s.., el.w. ..4 1,,

• C.,,.7 0.,

,,.,1 e O.,C.. ..,., r.1. w,. n,. ",..., -., of7 e r.,. -L.. ...v. ..m ~.. +...- c 4.w g., m.,... 4.. o. 4 s. . f n C.4... ...w..,,. a 23..4.. 4, v....,3. .....g... ,/*. :L, f.A.

n. s. a..

-.6...r ....a o .O..,,.a., [. 7., A .#. a.. a 4 .-m 1 4 *g. 1 4 w ,.,4 .g.4.mg... 4 .. ~... v.... .a. %....w w

1. . v,,

,..n.,,...,. ". m.. S, L. C, ...,... e S, 7.2 3.)... .i ^ *i w' m.. 0^ f .e .....w.

e. s.., w.i, n.,

.a w S. a a.,7 34... .X.4 a.g,..w,..s g... 9.., .O,.-a.. --.y. M. g 4.. a.w v ...g-gs .s ,nt va an, .0, ofev. ...w m 4 (',. r.3,.. w,,g......., 3. . w...,2.e....., '..., r...

g... g w..C v*.'. r a. a.".".4

.a .w., ...g .....eg .....e ...., _',,, C'.L,A. A., * ^ cf. S... u. ,e .w. ..,. -.n a ....3.. w. i,. O. a.s,7 S. 3 .w ..a....., .,, s o u. 4"w S * *..*.S.ka"3, '.. d., Y.X* a.." #. a..". *- .#*' * * #.". a d. ^ * ' a " '..m. *. #.".e'

• . k a.

Y..*# y g f .w y aX.4, yv a a.4 k..m.

4.,..,,,. v. a-4

.w. ...r.. ....... 4v J'13.- .3 4.,., 4,. c.~..a.. or w. v. g -., ":ae" . 3.e..u g.., .G. a. 4.e. 4 s - . y,1~~. ..x. .?... 4.+ a o,. w w e.i.4 -, 4.., s4 0, o.

4... a.., a w.

y,a.,. .,,. ---.4.,

4.c '.,.

c.4 ..v~. s ..a .......w, n &..,, 4. a, i.e.. ,e . 4.

4..s -..

1 s.y.. ..g.,...w i 4.

4. w...

..~tj .,p*, -a,~=

e,.

f,. y 3 A.W 2,.gg.. , D.,dv... N i.. .23 4.,.a.. 4.,2. 4 3,n S a a..&.* g,. 4 - .,.1.,. 4.3 a.a c ..<p.3, S.4.. _34..,a. ,a-.4 .eS,

,,,,~.1 c. -. - -.

.c.m,. m. r.1 4e s ^Mn;..'. ::: .s c"13 O j,,

v.. g,.1 o.? p,. 5d., S a v....,

c.s 4S. 4.., &.* ,,. 4 4 C.,. 4 -,. O. .w. f... .v -o, w 1 . ~,, u. v, e a.X.4 3g. 4... a. e .w ...9

4. 5

.t, a n*t * !,, c,. 4..S**.. -vs. c m,, an."., ys .....r.... .,,t .y1

4. a.,..w.o,., V3.,

A. v., c- ...e e, i a 4' 734. - ,3. 3 4 4 ww~ .. ;.. 3. .v. ..4., .. 5 ."SX rs ...I. a.., 4.

l. 4.r. oc q,4. c-

4...,. A. 4.

4. O. a., "D.44.. &",

v4.a..4.%. 8./, o'f a **. .v. v. ....a w ^~' a

• ~, r. 4

,./. g A. "....r.e. 7 r.., a., y

3..,

.,,. 4 4,, 4 4 ....g . e. . 4. p. n., v. .w 4 .v w. a..a a.,.

4.. 4....

ww a.e. 4 - . u.p. or. w. 3.,. w...r..m'-*b.,-- .a 4 i-~4o-7 S - r..J 3 - o* .r w., 4 .a v. g v.. +- e 1.... e, O. .w. r.,

4. w a.4 f.3.-..3. %..., 4.. D.r,b3e S 4.,,..w.

n.,,,.

4.. a. 4...

,4 y s. _ w_... ~e.e Ja .,y.., '.*4 a, i '.<. a ", .V..,., S %,..,., a ca.... n g.e. e a.e-.. +-..., s. n 1^y. 7 t 1 ........nwe,.. r ~**."..*-s---~~mmm=v~ ~ a aw' ~ w ^*

68 APPENDIX prceress perort: New Enziand Crust and U: er Mantle Structure The recent establish =ent of the northeastern seismic array has allowed us to construct a preli=inary =edel of the crust and upper santle structure beneath New England. Because the array has only been in full operation for approximately 2 years, the dataset is limited, and we have analyzed the data using a variety of techniques including: o i 1. observations of relative J3 residuals 2. a ti=e ter= analysis using ? arrivals 3 three-di=ensional modeling using teleseisnie P-vaves h. analysis of array diagrams 5 refraction studies Preliminary results indicate a crustal thickening under central New Ha:pshire coupled with a slight crustal thickening vestvard towards the North A=erican craten. Thero is also sc=e suggestion of a regien of relatively lov velocity in the upper tantle beneath central New Hs=pshire and southern Maine. Metheds of Analysis and Fesults The relative arrival tt=es of teleseisnic ? vaves were read fres 4 enlarged copies of 16 =m develocorder film. In general, the first few cycles exhibit coherence acrcss the array so relative arrival =easurements vere taken frc a preminent peak or trough early in the signal. This-procedure was required fer a nu=ber cf weakly recorded teleseis=s in I which the first break was too emergent or obscured by noise. In this way, arrival tt es could be =easured to 0.1 sec. Elevation corrections were applied to the data by assuming a vertical phase velocity of 6 0 l k=/see and dividing this into the station elevations. l l

69 Absolute travel ti e residuals vere calculated with respect to J3 tables and are defined tc te .s 44 v.s .n. o.

4. >4

.3 4 au J*~ is the absciute residual with respect vhere 3 to J3 tables for 4J station i, event j; T,o*~s is the Observed travel ti=e using origin 4 .a ' d _-a s f-.m 2."..? '. "~' 1 a. *. 4.. s, *,, * ~ d s +.".a..'..a.o r a. w ' al +. ave' *.* e *.*. - ugh. a JE earth. The residuals were reduced by calculating relative residuals with respect to a =ean residual ec=puted for each event; 'l

• 2 J2 J. p4 w

44 43 ma u 4., Where N is the nunber of staticns reporting ? arrivals for a given event. The utilization of relative residuals reduces source effects and =islocatien errers, re=cres errors in origin ti=e, and reduces effects of travel path thrcugh an inhc=cgeneous =actle. In this way, positive t i residuals represent late arrivals where the vsves have been slowed in i the crust er upper tantle beneath the array, i There are seversi censistent trends in the teleseis=ic P vave J t 1 i residuals which sugges* the presence of large scale regional structures in the crust and upper cantle teneath the array. The data show both a inuthal variaticas in residual values, and variations in average i station residuals across the array. 4 _...... a..... 4..,.w,e._.. 2,:'s kn ".s ' - 'e'..a. ."..-a.=. d.4_a i -w.. 4,._..._. /- s.:,s s.,.., _. a... 4. ...,w.4-... ,4 4 ..._...s.

..wo-

.-* in.e._ ...--,s.w. _ 2 .~. l l esting resu1* is the presence of a regional :cne of relatively Icv i l velocity in the upper tantle 'ceneath central 5ev Hs=pshire and southern-f Maine. This zene of relatively low velocity correlates spatially with 1 . h.g....I'.O 4A4g . 6 4 9. g .I I._% t.e.r.v. 4 D.e

• 4 4,.9 G.4.4 76 4 4w y p h..

pg hh .wg l

• g 4M i

w s. .w f. .w ,w.. i y 9

• 99

+

70 ser:e of these intrusive c0=plexes is deep-seated (Chap =an,1976), and it is possible that this anc=aly is related to the for=ation of these plutons. S ti=e ter= analysis using ? arrivals indicates that the variations in average station residuals =ay be due to variations in crustal thickness o and/or velocity. This is in centrast to the observed a i=uthal distribu-tion of residuals for each station which is probably due to deeper e effects. It was assu=ed that the distribution of average residuals is t caused by crustal thickness variations, and the data vere inverted to find a crustal thickness map of Nev England. The resulting =ap suggests a crustal thickening beneath central New Ha=pshire, with =cre normal thicknesses in Massachusetts an:1 Maine. The conteurs of the =ap parallel the northeasterly trend of the Appalachians. ~'he variations in crustal thickness observed across the network are also supper:ed by analysis of array diagrs=s. '~nese are stereographic projections of sicvness and azi=uth anc=alies observed frc= a plane wave fit to the wavefront traversing the network. These studies indicate a Moho which dips 2 or less to the n0rthwest. This is not surprising because it is expected that the crust vould thicken frc= the centinental =argin towards the Porth A=erican craton. In addition to the above =entioned studies, an average ernstal velocity =edel has been c:= piled for eastern Massachusetts and southern New Earphire by ec=bining results from ti=ed quarry blasts with the time term analysis. Tne =edel is er rently being used in earthquake location prograss at M.I.T. and is as fc11ovs: Gt a 4

71 laver (k=) F velceitv (kn/see) 0-73 5.65 c.-c. c s. --.- 9 26.1-33.0 7.33 Moho S.13 Future Studies e Studies for the next year vill be sined at inproving the preltninary crust and upper =antle =cdel for Uev England. This vill be achieved by using additional teleseismic ? and Fk? data. The database is currently being expandei to include readings frc= shcrt period stations in Ocnnecticut and eastern New York. The structural models derived frc= the residual studies vill be ec= pared to these fret 1 ng period surface vave dispersica studies. Phase velocities are presently being ec=puted as a function of azLnuth frc= the Quebec-Maine border event of June 15,: 1973, and si=ple crustal .=cdels will be developed. Phase velocities villLals2 be =easured using the tvc station techniquo Mere elaborate =cds a vill be generated by perfor=ing a si=ultaneous inversion of phase velocity and attenuarien folleving the techniques cf Lee and Scic=cn (1975). A study of the lg phase, a shcr period higher =cde love wave, vill be initiaced to ec= pare the effect cf regicnal geologic structure on Lg propagation. The data vill te collected using three ec=penent, digital reccrding event detectors developed at MIT.

72 ?eterences Aki, K., A. Christofferssen, and E. S. Husebye, Determination of the three-dimensional seis=ic structure of the lithosphere,1. Geophys. Res, 82, 277-296, 1977 c Cha;=an, C. A., Structural evolution of the '4hite Mountain -ag-a series, Geol. Soc. M, Mem., lh6, 281-300, 1976, c Lee, '4. B. and S. C. Solomon, Inversion schemes for surface wave attenua-tion and Q in the crust and the mantle, Geophys. 1. R. Astron. Soc., M,47-71,1975 p i C 4 1 I i t

71 layer (km) P'velceity (k=/see) 0 - 7.3 5.68 ,-c-g o.cc 26.1-12.0 T.33 Moho 8.13 Future Studies Studies for the ne.n year vill be ai=ed at i= proving the prelbinary crust and upper mantle =cdel for New England. This vill be achieved by using additional teleseismic ? and Pk? data. The detabase is currently being expanded to include readings frem shcrt period stations in Ccnnecticut and eastern New York. The struct ral =cdels derived fres the residual studies 411 be compared to these frc= long period surface wave dispersion studies. Phase velocities are presently being ec=puted as a function of asi=uth frem the Quebec-Maine border event of June 15, 1973, and simple crustal codels will be developed. Phase velocities vill also be censured using the tvc station technique. i l More elaberate =cdels vill be generated by perforsing a simultaneous inversion of phase velocity and attenuation folleving the techniques of Lee and Sole =on (1975). A study of the Lg phase, a short period higher =cde Love wave, vill be initiated to ec pare the effect cf regional geologic structure en Lg propagation. The data vill be collected using three cc penent, digital recording event detectors developed at MIT. i l

I l 72 References

Aki, K., A. Christoffersson, and E. S. Husebye, Deter =ina-ion of the three-di=ensional seis=ic structure of the lithosphere, J.. Geophys.

Res, g, 277-296, 1977. Chap =an, C. A., Structural evolution of the 'a*hite Mountain -ara series, Geol. Soc. A,, Me=., lh6, 281-300, 1976. Lee,'4. 3. and S. C. Solo =on, Inversion schemes for surface wave attenua-tion and Q in the crust and the =antle, Geophys. J. R. Astron. Soc., M,47-71,1975 i ? 0 6 I b

e 5 J

• • e t

Y a i s P n :hlbit 4 M e i 5 e a

f w r~L.%c:- 6..eD P.pgzgever a na.n ame .-..m. Ilane : Michael A. Chinnery-Date of Birth: 27 Septedber, 1933 j Place cf Birth: London, England i 4 Citizenship: U.Se i 1 i 'arital Status: Married, two daughters Wife's Nane: - Thora Elisabeth (nee Hawkey) Wife': Place of Birth: Peterborough, Ontario, Canada Wife's Nationality: Canadian 1 l Military Service: Royal Air Force, 1952-5k 6 Rank: Pilot Officer (Flying Officer, Reserve) 3 ranch: Fighter Centrol i Area: England an* Igy;; Present Hene Address: 110 Gray Street Arlingten, Massachusetts 0217h U.S.A. Presen: Susiness Adiress: 1 i I Lincoln Laboratory, Me!.T. h2 Carleton Street Cadbridge, Massachusetts 021L2 U.S.A. Telephene: Ecne 617/6h6-0937 Susiness 617/253-7852 Security Clearance: Secret t t / / o i .1 h 6 9..' er.. ~ e ee eo e e e o eseeoe e eoe oe oe eeoe eeeo Q a J s. _w e. e.ee. ,,e o e o ee ee e eoeeeoeoe- * - e.y a J l )

4 1 EDUCATION High School: Brentwood School, Brentwood, Essex, England 1944-51 I Praepostor 1949 School Praepostor 1951 Head of House 1951 County Major Scholarship 1950 i State Scholarship 1951 1 Undergraduate: Corpus Christi College, Camb_Idge University 1954-57 Mawson Open Scholarship 1954 Caldwell Open Scholarship 1955 Foundation Scholarship 1956 Natural Sciences Tripos Part I 1956 (1st Class) Natural Sciences Tripos Part II (Ehysics) 1957 (upper 2nd Class) Graduate: Geophysics Laboratory, Department of Physics, University of Toronto, Canada 1958-62 Computation Center Fellowship 1959 1 Canadian Kodak Fellowship 1960 Imperial Oil Fellowship 1960 National Research Council Fellowsliip 1960 M. A. Thesis: "The Application of Dis!ccation Theory to Geodynamics" (Advisor: J. A. Steketee) Ph. D. Thesis: "The Dynamics of the Strike-Stip Fault" ( (Advisor: F. S. Grant) t Degrees: A.V.C.M. 1945 Victoria College of Music B. A. 1957 Cambridge University M.A. 1959 University of Toronto M.A. 1961 Cambridge University Ph. D. 1962 University of Toronto i M. A. (ad eundem) 1967 Brown University Sc. D. 1977 Cambridge University J p.-e- ,, + .zw e--r e - - - - - *

• v w e-vr--

--r

• 1. ^

EMPLOYMENT ~ Trainee Engineer (computer construction and development) Plessey Company, liford, Essex, England 1954 (summer) Geophysicist (seismic exploration and interpretation) Seismograph Service (England) Ltd, Keston, Kent 1957-58 Research Assistant (operate mass spectrometer) Dept. of Physics, University of Toronto, Canada 195S (summer) Geophysicist (field part*/ leader; seismic, magnetic, and electromagnetic exploration) Huntec Ltd, Toronto, Canada 1959 (summer) Lecturer (part-time) Dept. of Physics, University of Toronto, Canada 1961-62 Instructor II Dept. of Geophysics, University of British Columbia 1962-63 Assistant Professor Dept. of Geophysics, University of British Columbia 1963-65 Research Associate 5 Dept. of Earth and Planetary Sciences. M.I.T. 1965-66 Associate Professor Dept. of Geological Sciences, Brown University Providence, Rhede Island 1966-71 i Professor Dept. of Geological Sciences, Brodm University 1971-73 Senior Research Associate Dept. of Each and Planetary Sciences, M.I.T. 1973-present Group Leader Applied Seismology Group, Lincoln Laboratory, M.I.T. Cambridge, Massachusetts 1973-present l l l

. - ~.. i CONSULTANT TO Huntec Ltd, Toronto 1958-65 Arthur D. Little, Inc, Cambridge, Massachusetts 1966-73 Earth Sciences Research, Inc. Cambridge, Mass. 1969-73 Lincoln Laboratory, M. I.T. 1971-73 National Aeronautics and Space Administration 1976-present l (Lincoln Laboratory does not allow its employees to consult for Industry) FIELD WORK Seismic exploration, Milford Haven, England 1957 Electromagnetic, resistivity, magnetic and seismic exploration in AlasPa, Northwest Territories, and Alberta 1959 Gravity survey, British Columbia 1960 i s Shallow seismic exploration, Northern Quebec 1961 Gravity survey, Northern Ontario 1962 Shallow seismic exploration, British Columbia 1964 l COURSES TAUGHT I Applied Geophysics (physics undergraduates) Applied Geophysics (geology undergraduates) 7 l Elasticity Theor'/ (graduate) Dislocation Theory (graduate) Introduction to Geophysics (undergraduate / graduate) Introduction to Seismology (graduate) i Earthquakes (introductory undergraduate) r [ Planetary Physics (undergraduate) Data Analysis (graduate) i l Tectonophysics (graduate) plus various seminars and portions of courses I

l t t PROFESSIONAL SOCIETIES AND OFFICES HELD American Geophysical Union Secretary, Tectonophysics Section, 1968-70 Program Chairman, Tectonophysics Section,1969 Annual Meeting Program Chairman, Tectonophysics Section,1970 Annual Meeting Associate Editor, Journal of Geophysical Research, 1969-72 Associate Editor, Geophysical Research Letters, 1974-76 Member, Committee on Education and Human Resources,1979-present Secretary, Seismology Section,1980-present Seismological Society of America Nominations Committee,1974 Seismological Society of America (Eastern Section) Resolutions Committee,1973 Chairman, Executive Committee, 1973-75 Member, Executive Committee, 1975-77 ' Society of Exploration Geophysicists Membership Committee, 1963-65

• Royal Astronomical Society of Canada Secretary, Vancouver Center,1963 1

President, Vancouver Center,1964 American Association for the Advancement of Science

• Society of the Sigma Xi Member, 1966-73 Treasurer, Brown University Chapter, 1968-72 Royal Astroncmical Society Fellow,1973-present 1

I l l

• presently inactive E

COMMITTEES AND MISCELLANEOUS ACTIVITIES Resident Faculty Advisor, Acadia Residence, University of British Columbla, 1962-64 Member, Gravity Sub-committee, National Research Council of Canada, 1964-65 Associate Resident Fellow, Mead House, Brown University, 1967-69 Member, Dining Services Committee, Brown University, 1969-71 Member, Graduate Council, Brown University, 1969-71 Chairman, University Lecture 6ips Committee, Brown University, 1971-73 Department of Geological Sciences, Brown University; committee memberships during the period 1966-73: Foreign language committee (chairman) Geology Club (chairman) Graduate examinations committee (chairman) l Lecture series (chairman) i Geophysics committee (chairman) l Undergraduate program committee (member) i Graduate admissions and awards committee (chairman) Testified.1efore the Advisory Committee on Reactor Safeguards, Nuclear Regulator'/ 'l Commission, concerning seismic risk at the Seabrook nuclear power plant il site,1974 Apps ared as expert witness at the licensing hearings for the Seabrook nuclear power plant,1975 l5 Member, Panel on Seismograph Networks, Committee on Seismology, National l Academy of Sciences, 1975-77 l l. Participant, Conference on earthquake prediction on the global scale, U.S. Geological Survey, Denver,1976 Chairman, Advisory Committee on Earth Dynamics, N. A.S. A.,1976-77 l Meeting Chairman, Summer workshop on the application of space techniques to geodynamics, N. A.S. A., Denver,1977 t l Member, Working Group on Upgrading WWSSN Stations, National Academy of Sciences,1977

}

i ; ll Gave specialinvited lecture on the application of space techniques to geodynamics, ji International Association of Seismology and Physics of the Earth's Interior, Durham, England,1977 ij" Member, Panel on Storage of Digital Seismic Data, Committee on Seismology, l ; National Academy of Sciences, 1977-78 l i

Member, Working Group on Solid Earth Data,.Committea on Data Interchange and Data Centers, National Academy of Sciences, 1977-78 Member, Proposal review panel, N. A.S. A.,1978 h Chairman, Advisory Committee on Geology and Geophysics, N. A.S. A.,1978-present Member, Space and Terrestrial Applications Advisory Committee, N. A.S. A., 1978-present Participant, Conference on Seismic Gaps, U.S. Geological Survey, Boston,1978 I. A.S. P. E. I. representative to joint I. U. G. G. /I. U. G. S. working group to formulate a post geodynamics program for the 1980's, Washington,1978 Member, Group of Experts study of seismicity in the Eastern U.S., Nuclear Regulatory Commission,1978-present Member, Proposal review panel, N. A.S. A.,1979 Participant, Conference on the Determination of Earthquake Parameters, U.S. Geological Survey, Denver,1979 Member, Seismic Research Review Panel, Vela Seismological Center, U.S. Air Force,1979 Chairman, Study on Geophysical Data Policy, _ Geophysics Research Board, National Academy of Sciences,1979-present I.C.G. delegate to Symposium on Quantitative Methods of Assessing Plate Motions, I. U. G. G., Canberra, Australia,1979 Gave technical presentation to the Nuclear Regulatory Commission on the application of probabalistic methods to the estimation of seismic risk, Washington,1930 Member, Panel on Data Problems in Seismology, Committee en Seismology, National Academy of Sciences, Woods Hole,1980 h ni m i u i

- -~ i a PUBLICATIO!IS The following list includes a variety of different kinds of publications. Papers in scientific journals are indicated by an asterisk (*). 4 t 1. Chinnery, M.A., The application of dislocation theory to geodynamics, M. A. thesis, University of Toronto, 88pp.,1959 3 2. j

• Chinnery, M. A., Sc=e physisal aspects of earthquake techanist, J.Geophys.Res.,55,2352-5h, 1960. '

3

• Chinnery, M.A.,-Terrain corrections for airborne gravity gradient

=easurenents, Geophysics, 26 L30-89, 1961. I k.

• Chinnery, M. A., The defor=ation of the ground around surface f aults,

Bull. Seis=. Soc.'A=., }_1,, 355-72, 1961. I i 5 Chinnery, M. A., The dynanics of the strike-slip fault, Ph.D. thesis, University of Toronto, 138pp., 1962. 6.

• Chinnery, M.A., The stress changes that acconpany strike-slip faulting, Bull. Seisn. Soc. An., j3, 921-32, 1963.

7

• Chinnery, M.A., The strength of the earth's crust under horizontal shear stress, J. Geophys. Res., j9,, 2035-89, 196h.

8.

• Chinnery, M.A., The vertical displacements associated with trans-current faulting, J. Geophys." 3es., 70; h627-32,1965 9
• Grant, F.S., Gross, W.H., and Chinnery, M. A., The shape and thick-ness of an Archean greenstone belt by gravity methods, Can. J.

Earth Sci., 2, h18-2k, 1965 10.

• Chinnery, M.A., Secondary faultip.g I:

Theoretical aspects, Can. J. Earth Sci., 3, 163-Th, 1966. 11.

• Chinnery, M.A., Secondary faulting II: Geological aspects, 4

I Can. J. Earth Sci., 3, 175-90, 1966. 12. Tokso, M.!I., and Chinnery, M. A., Seistic travel times from -Long-shot and structure of the cantle (abstract), Trans. A=. Oecnhys. Union, _h7,16h,1966. j-13 Chinne 7, M.A., The dislocat' ion fault model with a variable discon- .uity (abstract ), Trans. A=. Geophys. Union, hl, 166, 1966. 14.

• Chinnery, M. A., and Toksoz, M.!!. - P-vave. velocities in the =antle belev 700 k=, Eull. Seism. oce'.

A=., 51,~199-226,1967 - 15

• Chinnery, M.A.,

Leonbruno,.W., McConnell, R.I., and O'Brien, J., Ctservations of fatigue deformation in contact loading,. ASME publication 67-02-53, 12pp., 1967 5

16. Chinnery, M. S., The vertical displacenents associated with transcurrent faulting, in Proceedines of the VE3IAC conference en the current status and future trocnosis for understandine the scurce c.e:hanisn of shallov seismic 1

events in the 3 to 5 naunitude range, VISIAC report 7835-1-X, 299-306, 1967 17.

• • okso:, M.

U., Chinnery, M. A., and Anderson, D. L., Inhomo-geneities in the earth's =antle, Geothys. J., 13., 31-59, 1967 18. Chinnery, M. A., Evidence for lateral variations in the lover =antle (abstract), Trans. Am. Geochys. Unien, h8,, 19h, 1967. 19 Chinnery, M. A., Source time function for a wrench fault move-ment (abstract), Trans. Am. Geothys. Union, h8,, 203, 196T. 20. Chinnery, M. A., Theoretical investigaticas of the mechanism of faulting, in U.S. Ucper Mantle Preiset Progress Recort, 126-7, 1967 21. "Chinnery, M. A., and Petrak, J. A., The dislocation fault model with a variable discontinuity, Tectonephysics, 1,513-29, 1968. 22. Chinnery, M. A., and Rodgers, D. A., The stressed zone at the lower edge of a strike-slip fault (abstract), Trans. An. Geophys. Union,hjt, 299, 1968. 23 Chinnery, M. A., Earthquake magnituie and source parameters (abstract), Earthouake Notes, 39, 13, 1968. 2h. Chinnery, M. A., Measurement of the first and second derivatives of the travel ti=e curve using LASA (abstract), Geol. Soc. Am. Special Pater 101, 29h, 1968. 25 Chinnery, M. A., Direct =easure=ent of the second derivative of the travel time curve (abstract), Geol. Soc. Am. Scecial Pater 115, 215, 1968. 26.

• Chinnery, M.

A., Velocity an0=alies in the lover mantle, Phys. Earth & Plan. Int., 2, 1-10, 1969 i 27

• Chinnery, M. A. Theoretical fault sciels, c ts. Dem. Obs. Ottava, u

37, 211-23, 1969 28. Chinnery, M. A., F.eview of " Hon-elastic Processes in the Mantle", edited by D. C. Tozer, Trans. An Geothys. Unicn, 50,, h97, 1969 29 Bodgers, D. A., and Chinnery, M. A., The displacements and strains associated with a curved strike-skip fault (abstract),Trans. An. Gecchys. Union, 50, 233, 1969

30. Chinnery, M. A., The velocity an zly at 2000 km depth (abstract ), Trans. Am. Cecchys. Union, _5_0, 2hh, 1969 31.

• Chinnery, M.

A., Earthquake r.agnitude and scurce parameters, Ball. Seis.. Soc. Am., 59, 1969-82, 1969 32. "Chinnery, M. A., Earthquakes and the Chandler votble, Cenrents on I Earth Sci.. Geophys., 1,, 1-7, 1970. 33 "Chinnery, M. A., Earthquake displace:ent fields, in Earthouske Disclacement Fields and the ? tatien of the Earth, edited by Mansinha and others, Reidel Press, The Netherlands, 17-38, 1970. 3k. "Chinnery, M. A., The Chandler vetble, in Understanding the Earth, edited by Gass and others, Arte=is Press, Great 3ritain, 89-95, 1971. 35 Rodgers, D. A., Chinnery, M. A., and McConnell, R. 't., An assess-ment of the glacial retcuni nechanism for earth; sakes in the Eastern U. S. (abstract),Trsns. An. Geophys. Union, 52., 277, 1971.

36. Chinnery, M. A., and Jovanovich,1. E., The effect of the lithosphere-asthenosphere tcuniary on earthquake displace-ment fields (abstract), TYans. Ar. Georhys. Union, 12, 3h3, 1971.

37 Chinnery, M. A., P-vave arrivals s: the Large Aperture Seisnie Array, in Unter Mantle Pre'e:t U.S. Fins 1 Ferort, 56, 1971.

38. Chinnery, M. A., Theoretical and field investigstiens of the mechanics of faulting, in U::er "antle Pro.!ect U.S. Fins 1 Retort, 16L, 1971.

Chinnery, M. A., Investigations of ?-vave arrivals at LASA, Final 39 Technical Report, ARPA Contrzet Fhh620-c8-C-0082, S5pp,1971. Wells, 7. J., and Chinnery, M. A., Variations in the annual ccaponent j LO. of polar =ction at indiviinal observatories (abstract), Trans. An. Geechvs. Unien, 53,, 3h5, 1972, bl. Jovanovich, D. 3., and Chinnery, ". A., Evidence fer a cusp in the travel time curve at 35' (atstract), Trans. A=. Oecchys. 5,,'+52,,9no. t u.. t n., . j_ h2.

• Chinnery, M. A., and Wells, F.

J., On the correlation between earthquake occurrence and disturbances in the psth of the rotation pole, in ?ctatien cf the Earth, edited by Melchior l and Yuni, Reidel Publishing Co., The Uetherlands, 215-220, 4e. 3 1. ~, .. - ~. -

W u 9 h3. Rodgers, D. A., and Chinnery, ::. A., A revised velocity structu e for New England (abstract), Trans. ;. Geophys. Union, 53, L>2, .'C 7_. ~~~ ,s Lh.

• Rice, J. R., and Chinnery, . A., On the calculation of changes in the earth's inertial tens r due :: faulting, Gac;hys.

J., 21,29-90,1972. h5

• Chinnery, ::. A., and Jovanovich, J. 3., Effect of earth layering on earthquake displacement fields, 3u'1. Seism. Soc. A=.,

_62, ' 4_C,-1_',9, '9'2. 4 h6.

• Rodgers, D. A. and Chinnery, M. A., Stress accu:::ulation in the Transverse Ranges, Southern California,_i_n Proc. Conf. en i

Tectonie Proble=s of the San Andreas Fault Syste=, ed. R. L. j Kovach and A. Nur, Geol Sci., M, School of Earth Sciences, Stanford University, 70-79, 1973 h7

• Wells, F. J. and Chinnery, M. A., On the separation of the spectral cc=penents of polar =ction, C;eephys. J., Jh, 179-192, 1973 h8. Chi =ery, M. A., Iarthquake risk in Scuthern New England (abstract)

Iartheuske Notes, _hh, 29, 1973 h9 Chi =ery, M. A., and Rice, J. R., Cn the calculation of changes in the earth's inertia tenser due to faulting (abstract), Geophys. J., 3_5_, 373, 1973 50.

• Chi =ery, :. A. and Roigers, D. A., Earthquake statistics in Souther New England, Zurtheuake Notes, d, 89-103, 1973 51.

Chinnery, M. A. (Iditor), Seistic 21scrimination, 20th Semi-A= ual Technical St-**y, Lincoln Laboratorf, M.I.T., 31 December 1973

52. Jovanovich, D. 3., Eussaini, !!. I., and "hi=ery, M. A., Displace-

=ent strains and tilt fields due to a point dislocation in a layered half-space (abstract), Ia:--hquake Notes, _h5,15, 197h. 53 Chinnery, M. A., The Inter.ational Seisni: 7:. nth: Introdnetion

__5, 2', 197h.

L (abstrnet), Zarthcuake Notes, Sh. Chi =ery, M. A. (Editor), Seis=ic Discri=i.:stion, 21st Semi-A=ual Technical S-ary, Lincoln Late: aterf, M.I.T., 30 June 197h. 55 Cni= ery, M. A., and North, R. G., Frequency-=agnitude curves and the M -n. relationship, in Seis i Dis:rinination, Se=i-Ar.nual s D i Technical Sun =ary, Lincoln laboratory, M.I.T., 30 June 197-., 1h-15 1 1

i a t 1

56. Chinnery, M.

A., and :forth, R. G.,The 7.0:ent-M, relationship, and the frequency of large earthquakes, h Seis=ic Discri=ination, L Semi-Annual Technical Surnary, Lin 01n *aboratory, M.I.T., 30 June 197k, 15-16. 57 Chinnery, M. A., A state =ent before the AI: Advisorf Cc==ittee on l Reactor Safeguards, October 31,197'., lipp. 58.

• Jovanovich, D.

3., Husseini, M. I., and Chinnery, M. A., Elastic disloca ; ions in a layered half-space - I. Basic theory and nu=erical =ethods, Ge w hys. J.* _3_9_, 205-213, 197h. 59 "Jcvanovich, D. 3., Husseini, M. I., and Chinnerf, M. A., Elastic disclocations in a layered half-space - II. The point source, Georhys. J. _3_9, 219-2h0, 197h. j

60. Chinnery, M. A. (Editor), Seis=ic Discri=ination, 22nd Se=1-Annual Technical S"-ary, Lincoln labora:Ory, M.I.T., 31 Dece=ber 197h.

61. Chinnerf, M. A., Characteristics of short-ter= variations in seis=ic l

activity,_in_ Seis=ic Discri=ination, Se=i-Annual Technical S"-ary, Lincoln Laboratorf, M.I.T., 31 Dece=ber 1974, L9

62. Landers, T. E., and Chinnerf, I!. A., Spectral analysis of earth-quake occurrence rates, M Seistic Ois:ritination, Semi-Annual Technical Su-w/, Lincoln *.sb0ra:Orf, '!.I.T., 31 Dece=ber 197h, h9-51.

1

63. Chinnery, M. A. and Landers, T. E., Short ter= variations in the level of global seis=ic activity (abstract), Geological Society of A= erica Abstract vith Pre rens, ~, 3, h01,1975 6k. Chinnery, M. A. and Landers, T.;I., Evidence fr = earthquake ti=e sequences for a large-scale event involving the Pacific and Nazca plates dring 196h-68 (abstract),.' ans. Am. Geoehyc.

Unica,j6,kh3,1975 65 Chinnerf, M. A. and North, R. G., Global frequency-sagnitude relationships and their i=plications (abstract), Earthcuake l Notes, L_6, 55, 1975 l 66. Landers, T. E., and Chinnery, M. A., Spectral analysis of earth-quake occurrence rates (abstract), Iarth:uake Notes, k_6_, i 55-6, 1975

67. Chinnery, M.

A., (Editor), Seis=i Discri $ nation 23rd Se=i-Annual. Technical Sc--*ry, Lincoln L2boratory, M.I.T., 30 June 1975 68. Chinnerf, M. A., Spatial and ten;cral variations in the frequency- =agnitude eu.rve, M Seis=1: Di:eri:1:ation, Semi-Annual Technical' S= ary, Lincoln La'ocrator/, ::.I.T., ~ 33 June 1975, 55 -m.

a i 69 Chinnery, M. A., and Landers, T. E., Correlations between seismic activity in videly separated regicns, in,Geisnic Discri=ina-I tion, Semi-Annual Technical Su==ary, Lincoln Laboratory, M.I.T., ~ j 30 June 1975, 56. f i 70.

• Chinnery, M.

A., The static deformatic of an earth with a fluid cere: I a physical approach, Cecchys. J., h2, L61 h76, 1975 71. Chinnery, M. A., and Landers, T. E. Short ter= variations in the 'evel of global seis=ic activity (abstract), Earthquake i Notes, L6, 26-7, 1975 I 72.

• Chinnery, M. A., and Landers, T. E., Evidence for earthquake triggering stress, Nature, 258, L90 h93, 1975

73. "Chinnery, M. A., and North, R. G., The frequency of very large earthquakes, Science, 190, 1197-8, 1975 I

Th. Chinnery, M. A., (Editor), Seis=ic Discrimination, 2hth Se=i-Annual Technical Su==ary, Lincoln Laboratery, M.I.T., 31 December 1975 75 Chinnery, M. A., Proble=s in magnitude esti:ation, in.Seic=ic Discriminatten, Semi-Annual Technical Su= ary, Lincoln Laboratory, M.I.T., 31 December 1975, 1-2. 1 v6. Christoffersson, L. A., Lacoss, R. T., and Chinnery, M. A., Statistical codels for =agnitude estination, in Seis=ic Discrimination, Semi-Annual Technical Su==ary, Lincoln Laboratory, M.I.T., 31 Dece=ter 1975, 2-5 77. Christoffersson, L. A., Lacoss, R. T., and Chinnery, M. A., Estination of network =agnitude and station detection parameters, in Seismic Discrimination, Semi-Annual Technical Summa y, Lincoln Laboratory, M.I.T., 31 Dece=ter 1975, 6-8. i ~ l

78. Ucrth, R.

G., and Chinnery, M. A., =b estination for large events in the PDE catalog, in Seismic Discrimination, Semi-Annuti Technical Su==ary, Linecin Laboratory, M.I.T., 31 Dececaer 1975, 11-12. I 79 Chinnery, M. A., Damaging earthquake probability studies in the Eastern U.S. and their potential application to nuclear power plant siting (abstract), deol. Soc. A=. Abstracts with Progra=s_, 8, 150-1, 1976.

80. Lacass, E.

T., Chinnery, M. A., and Christoffersson, L. A., Ele =ents of a new. statistical apprcach to magnitude esti=ation (abstract), i 15~, 235, 1976. EOS Trans. A=. Geophys. Union, 1 i l

81. Chinnery, M. A., Lacosa, lR. T., and Chrictoffersson, L. A., Esti=1 tion of =b by a network of U.S. seis=le stations (abstract), EOS Trans.

A=. Oeophys. Unicn, 57,, 285, 1976. l 82. Lander:, T. E. and Chinnery, M. A., Spectral analysis of earthquake i .tica series (abstract), Isrthe_uake Notes, L7, 17-18, 1976. r k

i ) i i d J

83. Chinnery, M. A. 'Iditor), Seis=ie Dis:ri=ination, 25th Se=i-Annual

• echnical S r=ary, Lincoln Labcratory, M.I.T., 30 June 1976.

4 i 8L. Chinnery, M. A., and :* orth, R. G., Cc=parisen of recent esti=ates of =agnitude bias, M Seis=i: Discri=ination, Semi-Annual Technical Su==sry, Lin:oln Laboratory, M.I.T., 30 June 1976, 7-9 t i l 85 Lacoss, R. T., and Chinnery, M. A., Seis=ic =agnitude si=ulation studies, e Seis=ic Discri=ination, Semi-Annual Technical Su=ary, Lincoln Laboratory, M.I.T., 30 June 1976, 9-13 i

86. Chinnery, M. A., and Lacoss, R. T., 7.agnitu:le differences between station pairs, h Seis=ie Dis:ri=inatien, Se=1-Annual Technical S"-a y, Lincoln Laboratory, M.I.T., 30 June 1976, lh-15.

i i i 87. '"tinnery, M. A., Nu=erical si=ulation of the magnitude capability of a seis=ic network (abstract), EOS Trans. A=. Geophys. Union, i J,8_,th3,1977 I

88. Chinnery, M.

A., (Editor), Seis=ic Discri=ination, 26th Se=i-Annual 2 Technical S"--= y, Lin:cln Lab:ratory, M.I.T., 31 March 1977.- 2 89 Chinnery, M. A., and Johnston, J. C., Isti=ation of station detection-characteristi:s frc= bulletin data, in Seis=ic Discri=ination, Semi-Annual Technical Su =a f, *.in:Oln Laboratory, M.I.T., 31 March 1977, hh-6. 1 90. Chinnerf, M. A.,. C =puter si=ulation of network perfor=ance, h ' i Seismic Discri=ination, Seni-Annual Technical Su=ary, Lincoln Laboratory, M.I.T., 31 March 1977, L6-7

91. Chinnery, M. A., Correlation of seis=ic activity with changes in the rate of rotation of the earth, -pn_ Seis=ic Discri=ination, Se=i-Annual Technical S".==ary, Lincoln Laboratory, M.I.T.,

31 March 1977, 99

92. Chinnery, M. A., Major proble=s of gecdyr.anies, Paper presented to NASA Iarth Dynemics Su==er Workshop, Soulder, Colorado, July 18-23, 1977.

93

• Chinnery, M.

A., Measure =ent of :._ vith a global network, Teetenorhysics, M, 139-lhk, 1973. 9L. Chinnery, M. A., (Editor), Sei:=i Discri=ination, 27th Se=i-Annual Technical 5"-m /, Lincoln Laborat0ry, M.I.T., 30 Septe=ber 1977 95 Chinnery, M. A., and Johnsten, J. C., Saturation of =b 8#*1** M Seis=le Dis:ri=ination, Semi-Annual Technical Su==ary, Lincoln Laboratory, M.I.T., 30 Septe= der 1977, 32-33. w-- m-- - - - -g a-- -qm--

= - - - - - - _ _-- -- - - - -- i i i 96. Chinnery, M. A., (Editor), Seis=ic Discrimination, 28th Semiannual Technical S"--'ry, Lincoln Laborat:r;, M.I.T., 31 March 1978. t 97. Chinnery, M. A., A study of =aximu= possible earthquakes, Annual Report, URC Contract ITRC-Oh-77-019, Lincoln Laboratory, M.I.T., 15 August 1973 (reproduced as "RC Publication NUREG/CR-0563). i i

93. Chinnery, M. A., (Editor), Seismic Discrimination, 29th Sesiannual I

Technical Su==ary, Lincoln Laboratoryi M.I.T., 30 September l 1973. 99 Johnsten, J. C., and Chinnery, M. A., Measurement of =b using SRO l data, in, Seismic Discrimination, Se=iannual Technical Su= mary, Lincoln Laboratory, M.I.T., 30 Septe=ber 1978, h6-7 100. Chinnery, M. A., (Editor), Seismie Discrimination, 30th Sesiannual Technical Sunnar, Lincoln Laboratory, M.I.T., 31 March 1979 / 101. Chinnery, M. A., Tcvards the elimination of bias in body vave l =agnitude, paper presented to the USGS Conference on Earth-quake Paraceters, Denver, CO, March 19-21, 1979 102. Chinnery, M. A., Scatter in observed r., values, _i_n. Seismic Discri-mination, Seriannual Technical 5":-ary, Lincoln Laborato:/, M.I.T., 31 March 1979, hS-9 [ 103.*Chinnery, M. A., A comparision of.the seis=icity of three regions of the Eastern U.S., Bull. Seis=. Soc. Ag,., j$l, 757-772, e 9:y. 10h. Chinnery, M. A., Seismic require =ents f:r a seis=ic data center, paper presented to US/ USSR /UK Cssprenensive Test Ban Treaty Technical Eelegations, M.I.T., August 10, 1979 I 105. Chinnery, M. A., (Editor), seis=le Discrimination 31st Sesiannual Technical Sw==ary, Lincoln Laboratory, M.I.T., 30 Septe=ter i 1979 1C6. Gann, A. G. and Chinnery, M. A., Seis=ic data center design, in, 7 l Sei =ic Discrimination, Semiannual Technical S"--ary, Lincoln Laboratory, M.I.T., 30 September 1979, 1-2. 107. Chinnery, M. A. (Editor), Design of a Seismic Data Center, cpecial. internal report to DARPA, 230pp., 30 Septe=ber 1979. 103.*Chinnery, M. A. and Gann, A. G., Advanced data =anage=ent and processing techniques for seis=ic applications, paper presented at the Internati:nal Symposiu= cn 15dern Computerized Methods-of Registration and Interpretation of Seismic Observations, Yalta, USSR, October 2h-30,1979 To be published in conference proceedings.

we 4a. ..s*A C 7.,4. f. m. ', .e 4c... .".4s ,.{ 4.. 4., * %.. a .iQ,Q. /.* k...t

17..

n. s ....) ..... s 2 m.e,w 44.a.3 .c... ?.4.

• .C.s..,....,

17

v..arCw. 1.9 s v^.

...s -.. g n. 2., =,=.'..o-.~.,~.'.=.~.'.~...=.~..~. s'.a*.us, ~ .'10. C'..'.....=. ~g, V.. n'. <. '. ^1.~., Q, .e. 4..... 3. ..,.k. 4. g7 ......7 , _.g s.e. 4 i g.4..., .c...s 4. w 4 g 3 a...- ..a n.$9.2a. e T.4.s. ,, s

• .s.C

..C , y as s .s,... s n'., v'~ =.. =.. a. 3. -. +' .a s... '. a. *. c * ~. ' ~. ~.' s '.. s '., =4 ".c, e.' s. - C *. 4.. *. ~,, ....Discrimination, Seria.nual "e:hnice.1 S;-m /, Lincoln 'aboraterf, .v.. .i., 9, ,a,..h .3 9,2.,. 2 ...e y..,e ~.4. e s + .f.g.4. g.4e or .w = a ak.... /.

a. m. e.w..,. 4...o,ev.

n., 11 o. v s V.. n. of CD 'setverk !!! using synthetic data, U. S. telegstion paper US/GSE/7 presented to the Cennittee on Disar-a ent, Geneva, 7 July 1950. Also presented to !!A?O Aivanced Study Institute en Identification of Seismic Sources, Oslo, IIcrvay, 818 Septen-w.,.,,930. 113. Chinnery, M. A. and Gann, A. G., resign and develo;=ent of a seis=le .e.. n ..3g.4.c. ,g .,t..: /e..a.se.'.ed +.o a..,....., v. .e.. v y.. 9.a 2.C. al. ..."<.sa...-. ?. w 7... a 4 .w. bw. 4.... a y. ev y.. .g ~..... s.%ya .e..c...s.y

g. 4.... e n

i..s..

4.o. 4 n gs. 4. 0.*. O ' =4 ..a se.e. e..g.gn.n 2 s.m Seis=ic Sources, Oslo, ::cr ay, 3-13 Se;tenber 1980. .}}