ML20148U027
| ML20148U027 | |
| Person / Time | |
|---|---|
| Site: | Seabrook  |
| Issue date: | 02/28/1981 |
| From: | Chinnery M NEW ENGLAND COALITION ON NUCLEAR POLLUTION |
| To: | |
| References | |
| NUDOCS 8103020414 | |
| Download: ML20148U027 (145) | |
Text
{{#Wiki_filter:. s - UNITED STATES OF AMERICA NUCLEAR REGULATORY COMMISSION BEFORE THE ATOMIC SAFETY AND LICENSING APPEAL BOARD
)
In the Matter cf )
)
PUBLIC SERVICE COMPANY OF ) Docket Nos. 50-443 NEW HAMPSHIRE, _et _al. ) 50-444 (Seabrook Station, Units 1 and 2) )
) i STATEMENT OF DR. MICHAEL CHINNERY ON REMAND TO THE ATOfiIC SAFETY AND LICENSING APPEAL BOARD SUBMITTED BY THE NEW ENGLAND COALITION ON NUCLEAR POLLUTION N ">
p / \
-5
; g#Qp ([ t s j FEB171 ggt >3 Othee of tra Secre'aff 6 Ug7 A [ Docketing & St.Tct
% Stanch 9 (D. o tcp bi>
3
/ /
810302 0h _
# s ' -
1
-1 i
l ONLTHE APPLICATION OF THE "PROBABILISTIC" METHOD' TO THE ESTIMATION OF SEISMIC RISK AT THE SEABROOK NUCLEAR POWER PLANT SITE by ( i Michael A. Chinnery Leader, Seismology Group , Lincoln Laboratory, M.I.T. ! 42 Carleton Street Cambridge, MA- 02142
- Phone: 617/253-7852 F
t February 1981 h e y-P't-t- ~
~w e += g g -g g y9.q v==e w,e yg y=-e.,,w-, y,y,vy,-,9,-,w-,y.7,g,,,, 7.,,_, p% .,.,,_g .,y , , ,,, , , . . .
r Introduction 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 regi.onal and local l ! geology and seismology and specific characteristics of local subsurface material" (NRC Rules and Regulations, Part 100, Appendix A, Section IIIc). y There are two methods which have been proposed for the estimation of the SSE: , (i) The " Deterministic" Method: In this case, the largest earthquake in the historical record in the tectonic province containing the site is taken to be the SSE. Some additional conserva-tism may be included by making'the SSE larger than the largest historical. earthquake, though l l this has to be based on geological evidence. In j the Eastern U.S. the l'ack of detailed correlation l 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 record is taken as only a sample of the long term l seismicity of the tectonic province, and an attempt is made to extrapolate this relatively short record to longer time intervals. In this case, the con-cept of the " maximum earthquake potential" used in l 1 h
. . . . . . . . . . . . _ , , _ _ , _ . . , . . _ , , . . _ . . . . _ _ . _ , _ . , , _ . _ . . . . . . , _ . _ , , . . _ , . _ . , , _ . . , _ , , . . . . _ . ~. ,_ _ , , . - . , _ . . . , . _ . . _ . ~ , , _ . _ -
.. ,-. .- . . . -. _ .- - ~ , - .-. . . , . ~. . . . - -
the definition of the SS'E has to be modified, and
^
the SSE must be defined'as that earthquake.which I will occur in the tectonic; province ~containing the , site with some fixed acceptable level of' annual-risk or probability. This acce'ptable level.of risk.is not defined in the NRC Rules and Regula-tions.
.The Nuclear Regulatory Commission has ruledL(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 approach to the Seabrook site.
The Historical Record ; In New England the historical record of earthquake occurrence is approximately 300 years long. The only catalog i 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 relatively few events since it was created, and can , contribute little to the assessment of seismic risk in the area. We are, therefore, forced to work with the historical data set, in spite of its inadequacies. Now we have to ask , two important questions: Supposing that we thoroughly I k
- - . . - . . . ~ . - . - . . . . - ~ _ . - -- --. . . - - - .---
, . . . . .e . . , understood the long term seismic characteristics of the ; area, how well can we predict the seismic activity during , I 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 compute future seismic risk. The second question cannot be disposed of so easily, I and lies at the heart of all controversy concerning the estimation of seismic risk. How can we use the historical record to make the most reasonable estimate of the long term seismic characteristics? In order 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. Selection of a Tectonic Province . The concept of a tectonic province is a legal one (a s defined in Appendix A), and has no clear scientific signifi-cance. The proposition that earthquakes must in some way be nabted to geological etructure and tectonics is inescapable, but it is not at all clear that large provinces can be defined e y w, - r e-, v- v,g, - e e gr -.w -w r ywe,-ef ee m geweid- e, w y .v.-r<--g-y , ce, m t-get wo-e , -g-orgwe y- gm w we -py. g.4,,gi,-,yyg. , m g.ip mg., pe r ,-iw g e es ,ra gipf __em,,,,,,i.,we,eet--- m -a r 6e ap.
4 within which the seismo-tectonic characteristics are in any sense uniform. Attempts to define such provinces usually lead to a wide range of interpretations (see, for example, McGuire 1977 and Tera Corp. Study, 1979). These difficulties.certainly 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 Hampshire and in Northeastern Massachusetts. In previous 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), and this is indicated by the broken line in Figure 1. 6 In what follows, we use this Boston-New Hampshire seismic zone as the tectonic province appropriate to the 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-cular area. Instrumental epicenters for 1975-79 (see l Figure 2) are roughly consistent with this choice; the l cluster of epicenters in Southern New Hampshire can still be seen, but recent seismicity near Cape Ann, Massachusetts, has been low (in apparent contradiction to the historical record). Certainly, neither the historical record nor the instrumental record lead to any good arguments for isolating
-.4---- s=,- - y
"~ ' ~- ~ . ..
j %.q o O o [ - o ./
)
1M 3 0 o h e%, i@o O g o ,\ . - 0 . - o ( O s o .- og
. O l o .Ogg .
* 'O
- O % '
, g O -
g o -
- 1. .
' , A' o o$ !
5 ,
- e ,
a 't 1 g g o , sarEN56TY WAGMaruoi oo i , .. s= n- {a f *s A (
~
o 0 \
'g ,,,.
GD I E
=--
.i . .f .
, i I
g o o E:
, =
i .. - . )I w__!E_, 43 O
"{*' ,.
M--E oo
- g kO -
. )
.f
,, g3 *~"~
Og o u [ s l e .
'I e en.e.<a eo , g
~
,,, ' k .- T-
- o. 43 3 i ,
O Og- o [ --- E-+
\,i,4'(s o =L,
, JL gg J i-z
~'.
o e EP"Cf MTRE (Vacenamiles 4 2 26)
~,
e W O *, . o g,,cg y7 p g gy,,,,,siaties yt 26) . 0 7FivE Epf CENTRES Q SE15uccaspg STATION h , ,
\ scat.Eeu uitEs l,
\
Fig. 1: Epicenters in Mew England from the catalog by Smith (1962, 1966). The broken line indicates the Boston-flew Hampshire seismic zone (feca Chinnery, 1979a). , _ _ . _ , , y - ,,- - - - - - - .-y- --y g y=, w , , y
00* 70' 76* 74* 72* 70* 68* 66* 64* j _ _ p 4 g.,
\
\/
; I.tcEHp .
- 47' I
\ . Single Event j' . n
. Felt Event ,. ,
)
7 i > *** 46* t .
/ .,, 45o
'\ -
- n., , , , ,
\.
x .
. ga y.
J u s.: ,4 .f l\'; %!: .
' 7 N . .#
,s . \
43o g
,g e, [.
3y 7.. . , , ai N ' 11 j,, [ G . h
/ ~p,g, ,e 42'-
.)
- T
~
\ * . \ t
- '/ ! . J
,\
\
/
/ ,- .-
< L
*- /' 4io
- -{9 .* ,k 1
**~'
l* g' { ., ,l , 49e _' b, ( . .
\
>{ /
1 .-.- ! .
>v.&..
1 . t , 39a M 1 . t'Ng ' i ,
-' \ '
~
38-L .. .
/ - .
/
Fig. 2: Instrtunentally determined epicenters durin$ the period October .
. 1975-December 1979 (from Bulletin No.17, seismicity of the Northeastern ~
- n 4 < oa emo unc 9,.wi 4,.o i,, , ininen uve n o non a,.. o tonns
the Seabrook site from seismicity in Southern New Hampshire and Northeastern Massachusetts. In view of 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 Hampshire 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 catalog. In addition, we use cumulative frequency-intensity counts, i.e., we count the number of earthquakes larger than or equal to a given intensity va'lue during a given period. The extraction of frequency-intensity data from a catalog l such as Smith's must be carried out with care, since the completeness of the catalog at lower intensities is likely to be a strong function of population density, and therefore of time. We use the approach described in Chinnery and Rodgers 1973 (Exhibit 1) here. Having extracte? and plotted the data for the Boston-New Hampshire seismic zone, we have three important question to consider: 1 -,c- y .p- - - - , r. , , . g -. p.
(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? (iii) is there some upper bound to the intensity of earth-quakes that can be expected in this seismic zone? , Let us consider each of these in turn,
- 1. Linearity of Frequency-Intensity Data Frequency-intensity data for the Boston-New Hampshire zone are shown in Figure 3 (taken from Chinnery 1979) (Exhibit 2).
Clearly, the data ars sparse. Por the period 1800-1959 only six data points are obtained (for intensities II to VII) and it seems likely that those for intensities II and III are un-reliable due to incompleteness (even though these points are based on the very recent period 1928-1959). The remaining four data points actually lie in a relatively good straight 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 interpretation is that the number of intensity VII events (3) during this period was unusually high, and that the intensity IV data set may be incomplete. If these comments are valid, perhaps only the intensity V and VI data points are at all reliable, and we can not make any conclusions from the data themselves about the linearity of the frequency-intensity relationship. In this case, we must r'ely on information f rom elsewhere.
--- v -
w vu- - , , --,+ e -.nn> --.,e,- e v. mv. ~w. x ,-, ,. ,- , + , ,
.1
. ) . .. . %,
- ..,,.:.., l o.s . .
i '
. BOSTON-NEW HAMP$ntRg . ,
1800 -1959 - , o - g L.og N i ." ' g = 2.15- C.59I
-0.5 - * . .
.m 4
+
y . . . .- a: - w ..
- a. -to .;. ..,,'
.I - e
- t. . " .z. - ;
9 . o o ,' .
.J
-ss --
1- % ,. .
' .,'. i ,' .
.- 2 0 -
.. ,s; 3: .
' . ' t E' I t i e . .
E H E n x . .
, ,j- .,.;.,
INTDJS3TY . .' '.;;
. .-^' ..
a.
-> -.,.; .s
. 1 Fis. 3:
Frequency intensity data fro:n the Easton-flew Hat:pshire seismic cene, derived from the Smith catales using the mathods given in Chinnery and Redgers (1973) (from Chinnery, 1979a). . e I e
i _lo_ In my view, the current situation can-be summarized as follows: The vast majority of' seismologists have accepted the linearity of frequency-magnitude data as a working i
, hypothesis. (See, for example, Evernden 1970, Veneziano 1975,
.) l and the references cited in those papers).. It'is, however, still a hypothesis, with no clearly developed theoretical l basis. And there are a few instances where non-linearities , I These have led to several publica-are apparent in.the data. tions proposing non-linear relationships, though in my view t these can generally be attributed to poor or inadequate data. , The linearity of frequency-intensity data has been dis-cussed much less. Several investigators have proposed linear relationships between-intensity and magnitude, (See, for example, Veneziano 1975) and, if these areLvalid, a linear frequency-magnitude relationship implies a linear frequency-intensity 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 1979) (Exhibit 2). One point should be made here. Intensity (i.e., maximum epicentral intensity) is a very different scale from magni- f
' tude, and the observed linearity in the relationship between , ' the two at commonly observed intensities has no sound theoret-ical basis. Certainly for very large earthquakes there must i
t
' be a departure from linearity, since intensity has an inherent t
upper bound (intensity XII) while magnitude is an open-ended Y '
.i _ , - , . _ , _ _ . , _ _ . . _ _ , . , . _ _ _ , , _ _ ,_
. 1 l
I scale. Note, however, that all sc. ales become unreliable for large events (roughly M>7), due to saturation and other l effects. In summary, the apparent linearity of much frequency-intensity data must be treated as.an empirical observation. Its wide acceptance by seismologists suggests that it is useful as a working hypothesis.
- 2. Slope of Frequency,-Intensity Data 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- !
cantly from region to region? The only study that has addressed this point is Chinnery 1979 (Exhibit 2). In that paper it was shown that there l seems to be a remarkable uniformity in the slopes determined from various areas of the Eastern U.S. Values of this slope 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 i of 0.57. i This is an important point for areas such as the Boston-j New Hampshire seismic zone, where some of the data points 1 may be unreliable. If we assume that the data are to be fit with a straight line with slope about 0.57, then we can use the most reliable data points (for intensities V and VI) to define the frequency-intensity relationship (see Figure 3). In my view, more complex relationships are not justified by the data. i ._
r > J 1 I l i 1
- 3. Existence of an Upper Bound' Intensity Having defined a frequency-intensity relationship, we would like to use this to extrapolate beyond the historical ;
data points, to give an estimate of long term seismicity. The question remains: How far may wa continue this extrapo - ,
-lation? Is there an upper limit to the size 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 know the answer to these questions at the present time. One aspect of the - problem is worth nentioning here. All seismologists (including the author) agree that earthquake size (however measured) cannot increase indefinitely. Physical constraints arising from the earthquake source mechanism will set a limit to both source dimensions and strain release. On a global scale, this upper bound is at a rather high level, somewhat above the largest known earthquakes. On a regional level, much less is known, and there is considerable disagreement between the-(guess-) estimates of different seismologists. In a recent study (Tera Corp. Study 1979), ten experts in the seismicity of the Eastern U.S. made estimates the largest epicentral intensity that might be expected in the Cape Ann, Massachusetts region. These are listed in Table 1, and illustrate the disagreement clearly. There is little point in averaging opinions such as these .
. Notice, however, that 5 of the 10 experts admit the possibility that the 'T ves --vse- ---y-awspe ay,, ,y __..,,,7 ,._,, qp,y ,,,,,m_ . _ _ , _
i l upper bound to earthquake size ma? be X or greater in this region. TABLE 1 Estimates of the Largest Earthquake 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 5 XII 10 VII VIII IX 13 IX X XI 7 6.2 6.4 6.7 8 6.0 9 5.7 6.2 6.7 11 6.0 6.5 7.0 12 5.75 6.25 l (Here, arabic numerals indicate magnitudes; as a rough conversion to intensities, 6.0 VIII and 7.0 IX or X.) In my view, the only valid conservative interpretation of this set of opinions is that we should admit the possibility
.of an intensity X earthquake in the Boston-New Ilampshire seismic zone, until convincing scientific evidence arises 1/
! that will persuade us to revise this value.
~1/ In its previous ruling, the Appeal Board indicated a difficulty in accepting that data from one area could be of any use in attempting to project the seismic charac-l teristics of another area. This problem is fully (cont'd on next page) y 9---. a--m ,
,,7-- ,gm- -
y - --
l I I 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 Hampshire to Northeastern Massachusetts. Following Appendix A (section V, para. a.l.ii), we assume ,that the largest earthquakes that can occur in this province will occur at the site. Frequency-intensity date 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 obtain a measure of long term seismicity. Extrapolation of the line is considered valid out to an intensity of about X. 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 ir.dicate 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/ 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 observation and is indecendent of the ~eological charac-tarisrics of the three areas. .
-1>-
(30svod tlaa Ha-mesmas
\
Sosm c Zoaa
\x .
O A-T A : 17do - f 9 f 7
.\ *
\
o - r Io - - [ a \ .
\
\ ~
.\ \
I
)
a . 1
~ ,
\
\'\ -
~*
. .. /..- ,
. .m..-. .,.-t
. . .t. . m :
g
,9' c;
...g!
3, ,.. K . T P~g - SEsT 3 R*r A . .
. :. ~ [O
. !o -
. - -. . . . y
- E'.; 5s. ..
f1 E b . . [k-[ : ,a.
. e.;r-
)ff-3-
J-.
\\ $ .
. . [',
e .- , c. - ' g . p ,
'l j o_^ - .
175f d.APG 'A W CBRTMQ uf&E - [0?
~ . :--
..m. . .:-
..~.-
u._.
, s, , , . - -.
. : -e
' C: -
~
.. :I2 !Y
-25 2 -
SER SRoew
?..- Mi JO SSc '
. . 4 4- +<
M *-
.r
\
." . .:- . .s:....:.
.s .s -
._.i
'
- s ; c..:. u-
-3
\\ \
s
. ~
~- ;-
fac!
- s 10 -
g ,--
\ si,eer c s7
. M y 0 50 0 42 ^\ \
\ ?
- s
\ \
l3 _ . , w _ . . - . !.1. .il, 1.L . ,. . . . _V LL. .1%. . . .. . f.iii. .. .._ C'.a
. . . .T .. :7xS m/
Fig. 4: Extrapolation of frequency-interisity data far the Boston-New
The open rectangle shows how this earthquake would plot on the present diagram. Clearly, that event is consistent withe our extrapolation from later data. The current Seabrook SSE of VIII is found to occur with an annual risk of about 10-2.5 (this corresponds to a
~
return period of about 300 years). An annual risk of 10 (return period of 1000 years) corresponds to an intensity IX, and an annual risk of 10
~
4 corresponds to an intensity of at least X. The problem that remains is 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 mentioned in the past, I am not aware of any formal definition of this risk, which clearly involves many societal, economic and political factors, Conclusics This case study of the application of the "probabilis-tic" method brings out all the main features of tha 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 earthquake. r+- v ve-y -w-ww-ywwow wow w --r - y- -y--+-pw="*"a + .---weto---- g - v
l l 1 References . Chinnery, M. A. and Rodgers, D. A., Earthquake statistics in Southern New England, Earthquake Notes, vol. 44, pp. 89-103, 1973 (Exhibit 3). Chinnery, M. A., A comparison of the seismicity of three regions of the Eastern U.S., Bulletin Seismological Society of America, vol. 69, pp. 757-772, 1979 (Exhibit 2). i Chinnery, M.A., A study of maximum possible earthquakes, in 1 Annual Report, NRC Contract NRC-04-77-019 (NRC Publica-I 'j tion NUREG/CR-056 3) , 72p., 1979b. (Exhibit 3 is the i same study as published by Lincoln Laboratory, MIT). Evernden, J. F., Study of regional seismicity and associated l, { , problems, Bulletin Seismological Society of America, vol. 60, pp. 393-446, 1970. 1 l McGuire, R. F., Effects of uncertainty in seismicity on es,ti-mates of seismic hazard for the East Coast of the United l States, Bulletin Seismological Society of America, vol. ,I 67, pp. 827-848, 1977. Smith, W. E. T., Earthquakes of Eastern Canada and adjacent areas 1534-1927, Publications o_f the Dominion Observatory, ll .
! Ottava, vol. 26, pp. 271-301, 1962.
l
1 i l Smith, W. E. T., Earthquakes of Eastern Canada and adjacent I i areas, 1928-1959, Publications o__f the Dominion Observa-tory, Ottawa, vol. 32, pp. 87-121, 1966. Tera Corporation, Seismic hazard analysis: solicitation of expert opinion, Report to Lawrence Livermore Laboratory, August 23, 1979, NUREG/CR-15 82, 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.
--me w o. y- g - .y -- ,, p-p -.
-m ye -e
1 1 .i F UNITED STATES OF AMERICA l ! NUCLEAR REGULATORY COMMISSION L l DEFORE THE ATOMIC SAFETY AND LICENSING APPEAL BOARD l l l l ) ~ In the Matter of ) l }
} PUBLIC SERVICE COMPANY OF ) Docket Nos. 50-443 j NEW HAMPSHIRE, et al. ) 50-444 j )
I (Seabrook Station, Units 1 ) and 2) )
)
l CERTIFICATE OF SERVICE I
' I hereby certify that copies of the " Statement of Dr. Michael Chinnery on Remand to the Atomic Safety and Licensing Appeal Board" were mailed postage pre-paid this 17th day of February 1981 to the i
following: t i l i
. Alan S. Rosenthal, Chairman *
- Dr. John H. Buck Atomic Safety & Licensing Atomic Safety & Licensing Appeal Board Appeal Board U.S. Nuclear Regulatory U.S. Nuclear Regulatory Commission Commission Washington, D.C. 20555 Washington, D.C. 205S5 Frank Wright, Esquire Assistant Attorney General Assistant Attorney General Environmental Protection Division Environmental Protection Office of the Attorney General Division State House Annex, Room 208 Office of the Attorney General Concord, New Hampshire 03301 One Ashburton Place ,
Boston, Massachusetts 02108 Thomas G. Dignan, Jr., Esquire Ropes & Gray Robert A. Backus, Esquire 225 Franklin Street O'Neill, Backus, Spielman , & Little Boston, Massachusetts 02210 136 Lowell Street Manchester, New Hampshire 03101
- Docketing and Service Section U.S. ':uclear Regulatory Commission Fay Lesr,', Esquire Washi'aton, D.C. 23555 Office sf F.:ecutive Legal Dirac Or .
U.S. Nuclear Regulatory Commission Washington, D.C. 20555
~
*Dr. W. Reed Johnson D. Pierre G. Cameron, Jr., Esq.
Atomi.c Safety & Licensin9 Gon :ral Cc.unsel Appeal Board Public Se rvice Co:"93nY o f U.S. *!uclear Regulatory Corr.iscion ::ew Han;oshire i Washington, D.C. 20555 1000 gl= street M a re.: h e .i . t c , :i': 03105 Ms. Elizabeth H. Weinhold
- Atomic Safety and Licensing 3 Godfrey Avenue Board Panel Hampton, New Hampshire 03842 U.S. Nuclear Regulatory Commiss:
Washington, D.C. 20555 I i
! Office of the Attorney General 208 S t a te House Annex Concord, New Hampshire 03301 d _ .. ' M .9 % .?',ct W1112,.aar'S . Jordan, III l
- hand-delivered 1
- - . . , , ,s. yw: , . . , . , . . . . . . v.-.g-
O # d @ e 9.Ip S 544
- M6%
4 e g Exhibit 1 f \ ID / / 4 DGGew
~ usse ,,
t t- FEB I 7 f93r> '4, ( %es O h aNr s ' l l , l i 1 l
----,-,w-,r-.- , , - -- -wvme-o,--w--rn..w-, ,-- e , ,-,,,-,c.- . , -. - + ,*,7g _.y,,--.- , , , ,.,,-9, ..,,,4 m.,,., gg,, pwy,,,wy.- e p. c , r,-
i 89 EARTHQUAXE NOTES, VOL. XLIV. NOS. 3-4, JULY DECEMBER, 1973 . Earthquake Statistics in Southern New England Michael A. Chinnery and Donald A. Rogers
' l Department of CooloBi cal Sciences, Brown University )
Providence, R.I. 02912
*d y .: s ABSTRACT: New England has the longest recorded history of earthquake '
activity in the United States. Because of the high population den- 7 I g/' 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 relacions in an at- .' tempt to estimate the mean return period as a function of earch- ' '
~
quake size. _ -e INTRODUCTION M, New England has the longest recorded history of earthquake activity ~ in the United States. Several catalogs of earthquakes in this area have been complied, 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 throughovt this repert. No attempt to include information after 1959 '* has been made, in order tc preserve the apparent homogeneity of Smith's ?) data set. - It saeas very likely that several analyses of these data have been made in the past. However, if this is so, the results of the studies are , not generally available in the scist tific literature. Instead, it is common to find, in reports by insurance companies, site investigators, city > planners, etc., vague statements conceruing 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 does not seem advisable to base major planning decisions on statements such as these. Instead, we must examine the historical record in consider-able detail. These data are far from perftet, but they are essentially all that we have. The level' of seismicity is low enough that little information can bo deduced from instrumental;ecords, which are only avail-able after about 1925. In addition, the his.corical data suggest that the seismicity observed in the last century. may, be unusually low. ' ~ The purpose of this paper, then, is t.o examine Smith's earthquake catalog in detail. We shall concentrate on 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 to , , of the earthquake data, construct recurrence relations, and attempt We shall e estimate the mean return period as a function of earthquake size. study both the area as a whole, and also several smaller subareas where much of the historien1 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 Salth's much larger diagram covering Eastern Canada and the Northeastern United States.
In all, Smith lists 729 earthquakes in the Northeastern United States. e ; Now at California Division of Mines and Ceology, Sacramento, Calif. 95814 i
.i+ pe ip ww.--rergy si v'er..wh-e -g w eie~-e-wememp,- --y u y --wm 49ey- f ,e +-----g- ,,y--..m -,y>,e- -gy=- ,>p,w ary-A*- ,9%we ' , - w e swwm .ee ,w7-,-av- vrp.
l 90 EARTHQUAKE NOTES i j - E b.q 0 o
/
O / -' O d' W ' e' o
- c. O OQ s O ? O i o g.
,'\
0 r~\'.'
'. o i
* % A'.. C h
.. .fs,Qg O
. o o o \ ;
..m..
Og- o
~~.k, o\\ o 5
- Q INf tNsifT g ntACNIfUOc
'} o o ,, .%. n-
*\ , {.
% \
- k- =
ep E-
.? e* II -
-f -F o
c, ,_ ;- OO / 3 a
*.0 o U--.:-
) ,f . .y e "*
4 O\ O
**ove =58 e oo E~~-~S o... . 8 3 r 7
. Q % O .I
~*
;O O '
\
*O QO " '
l
- h. 6 * - . g 5.- S
. --c~3 O*p< e
\ '
51 - e (P1CE NTRE IUacencialett e t 21) O t met N t a t (uaceae.asies a s 20'l 0)nvt (PictNints CD scisuconata sf A;6cw O 'O ,. M g .# w _ -- scatt em uitts i ! Fig. 1. Earthquake epicenters in the New England l area,1534-1959 (af ter Smith,1966). We have chosen to select a portion of the New Eng, land area for study. This portion, which we term " Southern New England," is shown in Fig. 2. Included are the states of Massachusetts, Connecticut, Rhode Island, and the southern parts of New Hampshire and Maine. The ocean tirea Easi: to l l l l e
,. a , y o y . -- v-v qy w--w w.w -- r y v"v*- sw r*"w++ ' " " ' " " * ' ' -
"- " " "*"-'" "- * *-"* " - ' ' " - ' ' " ' ^ ^ ^ ^-
s EARTHQUAKE NOTES,'VOL. XLIV NOS. 3-4, JULY-DECEMBER, 1973 91 sa 44 N
~ . B .
- 3. , ,
j,. o. e
, 3 fi 0
= _
43 >
'Q r
P A r (
% =_: 42 0
l 1 - C 5 p t-l 4 ff 41 73" 72* 719 70 W Fig. 2. The area that we have chosen to define as " Southern , New England" for the purposes of this study. 69.5 West has been added to these, so that a number of poorly determined epicenters off the coast of Massachusetts may be included in the statistics. ! New England (defined above) is more distant than 50 miles from a major cen-l . ter of population (100,000 inhabitants or more). 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 some importance in attempting to estimate earthquake risk in the area as a whole. Much of the seismic activity in this region (see Fig.1) is c-e centrated into three zones, which are labelled in Fig. 2 as A B, and C. A includes the area around Boston, 3 ref ers to the southern part of New Hampshire, and C denotes the region of Connecticut around Hartford. The division between zones A and B is rather arbitrary, and the statistics of these zones are analyzed, both separately and together, later in this ' report.
?
9 - --=-+m - ,-
- l 93 EARTHQUAXE NOTES, VOL. XLIV, NOS. 3-4, JULY-DECEMBER,1973 .
Table 1. Subdivision of data into subsets. Subset 1 Subset 2 Subset 3 Intensity 1800-1959 1860-1959
- IX 1700-1959 1700-1959 1800-1959 1860-1959 VIII 1700-1959 1800-1959 1960-1959 VII 1860-1959 VI 1800-1959 1800-1959
,e 1860-1959 1860-1959 1860-1959 V
i 1900-1959 1900-1959 IV 1900-1959 1928-1959 1928-1959 1928-1959 III 1928-1959 1928-1959 1928-1959 II nature of the historical data, we shall use intensities throughoue. It appears in general to be possible to relate the maximum epi-central intensity I to the local magnitude M by a lir. ear algebraic ex- - pression. Cutenberg and Richter (1956) dntermined the following rela-tion for Southern California: (1) M=1+fI. The number of earthquakes in the present area for which both M and I l are known is small. Figure 3 shows that these data are consistent with l the Cutenberg-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 may convert instrumental magnitudes into inte.sities (some recent earth-More im-quakes in Smith's catalog are listed with only a magnitude). portant, it enables os to compare our, local statistics with global studies, which are usually quoted in terms of magnitude. It will be convenient at this stage to define what we mean by a damaging earthquake. On the basis of the Modified Mercalli Scale of in-tensity, the onset of considerable destructive ability occurs at an epi-central intensity in the range VIII to IX. We have therefore chosen to define a damaging earthquake as one with an epicentral intensity of VIIII/2 or greater. This is a vcry conservative choice, since even an carthquake of intensity VII is likely to cause some damage, particularly if it occurs in a heavily populated area. An intensity of VII1 1/2 corresponds to a magnitude of about 6.5.
' The possible destructive effects of an earthquake of this size may be Consider-illustrated by the 1971 San Fernando earthquake in California.
able damage and rather high accelerations of the ground were observed in the case of this much publicized event. In addition to the size of an earthquake, we must also consider the areal extent of the region subject to damage. This is much less well defined quantity, since it depends so much on the Juperficial geology. Linehan (1970) has given an earthquake intensity attenuation
, y-w,- -
m - - - , - , - - , m , . , - , - -
92 EARTilQUAKE NOTES In all Smith lists 353 epicenters in Southern New Encland. 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 lucations, problems in the determina-tion in intensity, and incompleteness of the data set.
' The uncertainties in most of the epicenter locations shown in Fig. 1 ,
are quite large. The historical data vill 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 instrumental determina-tions of epicenters are also subject to error, though oi a different kind. The seismic travel time curve in New England is not veli known. This is - in part due to the heterogeneity of the geology, and in part'due to the poor spatial distribution of epicenters ir relation to observatories in the area. Even the large New Hampshirc carthquakes of 1940 (M = 5.8) cannot be located to better than +20 km. It is doubtful whether any of the epicenters on Fig. 1 are any more accurate than this, and most are much less accurate. The determination of the intensity of an earthquake 'rca SJstorical eyewitness accounts is notoriously difficult. Estimates are racjact to population distribution, the personal feelings of the observer, and ttw interpretation of the cataloger. The influence of these factors is mix 44. Observers are likely to overestimate the intensity of an earthquake shock. However, the cataloger has clearly tried to take this into account in his assignment of intensities. In addition, if the population density was sparse, it is quite possible that no report was received f rom the highest i intensity zone close to the epicenter. In view of this, it does not ap-pear 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 in, of course, its completeness. There is no doubt that the data becomes more incomplete as one goes further into the past (at a given intensity) and to smaller - intensities (at a given time). On the other hand, we need the longest time period and the largest range of intensities possible in order to arrive at meaningful statistics. Because of this, some subjectivity is necessary 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 carthquakes since then have been recorded. Similarly, it is only in the recent past that we may expect a fairly complete record of small events. Therefore, in order to try to exclude this type of deficiency from the data, we have chosen to analyze three subsets of each data set. Tnese subsets, showing j' the time interval studied at each intensity, are listed in Table 1. By comparing the results for the three subsets, we will, hopefully, obtain some information about the completeness of the dats set as a whole.
INTINSITIES AND MAGNIT'JDES Intensities mentioned 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 intenalty of the earthquake at the observing site. Magnitudes can only be determined reliably from instrumental records. Because of tha g-w s v v
2._... . .. -. . 9/, EARTHQUAKE NOTES Vill - / . Least Square Fit /
.,. T^
M = l 2 + 0 61
- 4 .#v yll -
/ im
's
/ , _w.
/ ..'.
~ yl . . .; .<.;.-
./
>-- / -
n .- :
- / -
m
/ -3 Z . . - . . :.
m N.
,. ** / r.
s
-r
= / ,e j ".
2, ly - .
/ ..
- 2 Gutenberg - Richter - r M=1+2I 3 til . 6 2 3 4 5
, M AGNITUDE M
'4 -
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 from an epicentral intensity of VIIII/2 This is also a conservative estimate, and the radius may easily be doubled in regions i of unfavorable geology. In view of this, and the high population density of the ares under consideration, it seems unlikely that an intensity VIIl /2 l earthquake could occur in Southern New England without causing considerable damage and loss of life.
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. I l l
)
1 1 i e
-,-w.-- wr-g , -, , . y .y w -e,._,. ._,,r_..
..-._.f__q,y.,,, g.,,..,-,wp;. ,,y __,.7.-.,.,,e,,,- ,.g,_,,,.,,,-.,,.,.,_.,,,,_,,...,y,,, ....-,.4-- n,, - . .
EARTHQUAKE N(TIES, VOL. XLIV, NOS. 3-4, JULY-DECEMBER, 1973 95 FREQUENCY-INTENSITY RELATIONS It has been clearly demonstratQ6 in many parts of the world that there is a linear relationship between earthquake frequency and earth-quake magnitude (see, f Jr example, rvernden,1970) of the following form: loE Ne = a - bH , (3)
~
where N is the number of eartnquakes occurring within a region in a given c with a magnitude greater than c equal to M. a and b are con-
' time period stants; a depends on the size of the area 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 nf the tectonic activity causing the earthquakes. Logarithms, un-less ,therwise stated, are to base 10. , If a linear relation exists between magnitude and intensity, as we have discussed earlier, then clearly we may write l (4) log Nc " * ~ dI' where, now, N e is the number of earthquakes occurring within a region in a given time interval with an intensity greater than or equal to I. c and d are constants. The Relation 4 is a very useful one. It enables us to use the data for smaller earthquakes, which are plentiful, to determine the frequency of occurrence of large earthquakes. In the . tater sections of this report I we shall attempt to determine the constants e and d from the data in l Smith's catalog. l l In considering the statistics of the earthquakes, instead of using the quantity Ne, it is more convenient to define the "mean recurrence time" (MRT). MRT is simply the average time between earthquakes with a given intensity I or greater, and is equal to 1/Netime periods. We shall be particularly concerned with the determination of MRT for damaging earth-l quakes. l Before we proceed, however, we must* consider the range of validity of Eq. 4 Where complete data has been obtained, the frequency-magnitude relt. tion (Eq. 3) has been shown to be valid over a remarkable range of marpitude (from greater than 8 down to less than C). Thore is some theo-ret.ical 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 limits are well beyond the sizes of the earthquakes that we shall consider in this report. We must next examine whether there is any evidence that there is an , I upper limit to the sizes of earthquakes to be expe*cted in the New England area. There seems to be some confusion on this point. In fact, there is j e no basis for suggesting that such an upper limit exists, and as ve shall l
- see, analysis of the historical data supports this statement. The largest ;
l' earthquakes that have been recorded in Southern New En61and are listed in l Table 2. At least one, and probably two earthquakes of intensity IX (mag- j nitude about 7) have biren recorded in the past 400 years. This length of record is far too short to conclude that an event with intensity X (or ' greater) has not occurred in the past ,or will not occur in the future. The Charleston, South Carolina, earthquake of 1886 had an intensity X, and occurred in an area that is somewhat less seismically active than New England. l l
d I
- l 1
96 .EART11 QUAKE NOTES a Large certh' quakes in Southern New England. . Table 2. . Date Location Intensity 1568 Rhode Island VII 1574 Rhode Island VII 1584 Rhode Island VII 1592 Rhode Island VII !
- July 11, 1638 off Cape Ann. Mass. VIII November 9, 1727- Near Newbury, Mass. IX ,
June 14, 1744 Off Cape Ann, Mass. 1/III l '
. November 18, 1755 "about.200 miles" East- IX-
s
, of Cape Ann, Mass. '
May 16 or 18, 1791 Near Moodus, Conn. VIII
- October 5, 1817 Northeastern Mass. VII j ,
December 20, 1940 Ossipee Lake, N.H. VII December 24, 1940 Ossipee Lake, N.H. VII , f We must therefore admit the probability that large earthquakes-will occur in South 2rn New England, if at infrequent intervals, until some new information arises that dismisses 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 complacency. It is conceivable that a long time has elapsed since the last large earthquake i in this area. If this were the case, the probability of one occurring in the near future could be quite high. RECURRENCE RELATIONS: SOUTHERN NEW ENGLAND We consider first the whole Southern New England region (defined in Fig. 3). Smith (1962, 1966) lists 353 events in'this area during the period 1534+1959, after all obvious aftershocks are removed from the data. The distribution of these earthquakes in intensity and time is shown in Table 3. Where 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. Intensity Before 1700- 1800- 1860- 1900- 1928-1700 1799 1859 1899 1927 1959 IX - 2 - - - - VIII 1 2 - - - - VII 4 - 1 - - 2 VI 1 2 3 1/2 1 1-1/2 ! . V 2 8 5 8-1/2 9 6-1/2 IV 5 16 24 13 22 21 III 2 16 40 23 14 26-1/2
' II 3 3 -
28 5 32-1/2 It is clear from Table 3 that the data from before 1700 are very incereple t e . At the lower intensity levels this incompleteness con-tinues until late in the historical record. For this reason, we have disregarded portions of the data, and have analyzed the tenainder in ' F I e i I I ' m , -. -- , . 4 . , . . _ _ . , , - - _ , ,+__-,,,.,..-,%... - , . . . , _ . . . - - , . , _ . . . . . . ~ . . ,,~ .-.~._ . . - . . . - , . . . - . . . -
4 97 ZAftTHQUAKE NOTES, VOL. XLIV, NOS. 3-4, .NLY-DECEMBER,1973 Southern - l 25
- New England 1 e
20 . Log .
- Ne 15 -
Log Ne : 030 -0 57I 7
. ~ l" 10 -
I n 05 -
}
t , 0- , l- vill X 11 IV vi INTENSITY I Fig. 4. Frequency-intensity plot for 135 events in Southern New Engla'nd. N e is the cumulative number of events , (with intensity I or greater) per century. Note, however, that all the three subsets described earlier (Table 1). These large earthquakes with intensity VIII or IX occurred before 1800. events will therefore only appear in subset 1 of the data. Frequency-intensity plots for the three subsets of the data have As may be expected, the large events in the 1700's been constructed. We conclude that subset 1 is unreliable. make subset i very nonlinear. identical, and we therefore have chosen to use Subsets 2 and 3 are almost The. subset 2 (which contains more events) as our most reliable The ordi- data set. frequency-intensity graph for this subset is shown in Fig. 4 note is the logarithm (to base 10) of the cumulative number of events with intensity I or greater, per century. The low points The points in Fig. 4 define a fairly linear relation. 1 is virtually im-l intensities !! and III are to be expected, sisce it
! at possible to obtain a compiece record of these small events, even in th 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 Theestimate data then determineof the slope of the frequency-intensity relation is 0.57 (to 03). 3 e
~"
ryYP'- y T*t$Ne t g"? un +y1's-++wgo gerp e e y w- 'p'm' 1'#w egprg'e+,ge*C-, g >pr. we g-m met dgiy MS t-E==#9,yMi'd rege *PM'M*eegy3-y e ww hw u.-N'wibt i v Weer - d-yme p$8 etwe g- wa -
'it'w-,'-me er e v"'see*w
4 98 EARTHQUAKE NOTES the following recurrence relation: Log Ng = 4.30 (30.15) - 0.57 (t0.03) I, (5) converting this into a frequency-magnitude relation; using Eq. 2, we obtain Log Ne = 3.45 (30.20) - 0.95 (10.05) M. (6)
/
The "b-value" in the range 0.9 - 1.0 is very reasonable for an area , such as Ncv England. b values lying in the range 0.8 - 1.0 are found in most parts of the world (Evernden,1970). Isacks and Oliver (1964) found a b valut 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 include contributions from errors in the data points themselves, which are extremely hard to estimate. RECUPJtENCE RELATIONS: BOSTON-NEW HAXPSHIRE RECION Areas A and B combined (see Fig. 2) include the Boston vicinity, Northeastern Massachusetts and the associated offshore region, and the Southern half of Nes: Hampshire. Suith lists 194 events in this active zone, which therefore accounts for just about 50% of the total activity in Southe rn 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 region (areas A and B combined). Intensity Before 1700- 1800- 1860- 1900- 1928-1700 1799 1859 1899 1927 1959 1X - 2 - - - - VIII 1 1 - - - - VII - - 1 - - 2 VI 1 2 4 1/2 1 - V 2 6 2 4 7 1 IV 4 6 '12 7-1/2 9 8-1/2 III 2 16 16 12 6 13-1/2 II 3 3 - 21 3 16 Frequency-intensity plots for these various subsets of these data a show very similar features to those found in the previous section for the whole of Southern New England. Inclusion of the early data (subset 1) leads to a very nonlinear plot. Subset 3 shows some scatter due to an insufficient number of events. Subset 2 again gives the most reliable data set, and the resulting plot is shown in Fig. 5. Linear relationships fitted to the data again have sicpes in the range 0.54 to 0.60. This (s an important point, for two reasons. Firstly, it substantiates the slope determined 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 resolution of the present data. Aa before, then, we assume a slope of 0.57 (10.03) for the fre-quency-intensity plot. This leads to the following recurrence relation:
EARTil0UAXE NOTES, VOL. XLIV, NOS. 3-4, JULY-DECDLBER,1973 99 t i i 25 . Boston - New Hampshire Region f 20 - Log Ne l5 - . log Nc = 4 00 - 0 57I f
. 1i 10 -
05 - i 0 - i - ; a a p it iv vi vili x INTENSITY I Fig. 5. Frequency-intensity plot for 65 events in areas A and B (see Fig. 2), which include the Boston vicinity and ' Southern New Hampshire. N eis the cumulative number of events (with intensity I or greater) per century.
, r e
- Log N = 4.00 (+0.15) - 0.57 (f0.03)-I. (7)
And, using Eq. 2, we find ; Log Ne = 5.15 (+0.20) - 0.95 (30.05) H. (8) i RECURRENCY RELATIONS: AREAS A, B, C Subdivision of the Boston-New Hampshire region into the individual
, subarcas A and D starts to point out soir.e of the inadequacies of the his-i torical data set. Superficially, Smith's catalogue includes 98-1/2 i e events in area A, and 95-1/2 events in area B. One is tempted to ascribe one-half of the activity in the Boston-New Hampshire region to area A, and
, one-half in aren B.
However, tabulation of the events in these two areas as functions of time and intensity shows up some oarked differences. Table 5 shows this L ,. ., ,,.,,,L-.~. ,,.-...,-,...-,,..-,A..-~_..-----. ..__...,_..._._-~.----.-.-.-~~-__.--m. - . - - -
'b i
100 ' EARTHQUAKE'NOTfS
- Table 5. Earthquake data for Boston' vicinity (area A). "
s Intensity Before 1700- 1800- 1860- 1900- 1928- , 1700 1799 14.59 1899 1927 1959 ! IX - 2 - - - -
, VIII 1 1 - - - -
yII - -
- 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 B). Before 1700- 1800- 1860- 1900- 1928-I"**"'I"I 1700 1799 1859 .1899 1927 1959 Ix - - VIII - - - - - VII - - - - - 2 VI - - - 1/2 1 - ' v - 1 1 2-1/2 2 1/2 W - 1 3 6 4 5 III - 3 12 9 1 12 II - - - 16 - 13 tabulation for area A,. and Table 6 shows the same for area B. Area A ap-pears to have had a relatively high activity in the 1700's, which has since been steadily decreasing. On the ochor hand, area B appears to show a low 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 have been heavily influenced by the effects of. population distribution, and.that the earlier parts of this
- data set are very incomplete. 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 uniform since the 1700's (see Table 4). Is this apparent uniformity real? It is worth while men-tioning the following possibilities:
(1) The area A data may be fairly reliable, while area 5 may be very incomplete. Addition of the " missing" New Hampshire events will bias all of our recurrence relations in the direction of increased seismic activity. If this is the case, we have a strong indication that the seismic activity ce ring the past 100 years or so has been unusually low. This may be the result of the statistical fluctuation, .or some unknown physical process. (ii) The early high activity in area A may be the result of exagger-ated intensity estimates for some of the events in the 1700's. If this is so, it is possible that the activity of the two areas has been relatively uniform throughout the historical period. g wv. . .,v..,. _.-,,,.-..m..-.. _,.--.....__.-...'.__._..---..~......_,,_.-_...---.....,_
)
i i EARTHQUAKE NOTES, VOL. XLIV NOS. 3-4, JULY-DECEMBER, 1973' 101 There is no way to distinguish between these possibilities using the present data. However, the second possibility will clearly lead to the , ! most conservativc estimates for the seinmic activity. Direct construction i
, of f requency-intensity plots 1cada to inconclusive results because of the )
small number of uneabic events. We therefore return to our first inclina- l tion,'and assume that the scismic activity of the Boston-New Itampshiro I region las evenly divided between areas A and U. This leads to the follow-ing estin.ates for the recurrence relations in areas A and 8:
' Log N, = 3.70 (+0.15) - 0.57 (10.03) I, (9)
Log Ne = 4.85 ,(10.20) - 0.95 (10.05) M. (10) e Clearly, 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 more reliable estimate 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 4.rea C. Smith lists a total of 55 events in , this area, of which only 20 fall in subset 1. This number is quite in-adequate for any statistical treatment. We may, however, get a rough idea of the activity in this region ' by assuming that the slope of the frequency-intensity is known (0.57 1 0.03), and that the record of events with intensity IV is complete during the period 1900-1959. .i This is sufficient to determine the following recurrence relations i for area C: ! Log N, - 3.35 (10.20) - 0.57 (10.03) I, (11)
, e .
Log N, = 4.50 (10.25) - 0.95 (10.05) M. (12) MEAN kECURRENCE TIMES From the recucrence relations listed in Eqs. 5 through 12, it is casy to cniculate the mean recurrence times. These are listed for a variety of intensities in Table 7. It should be noted that those were determined from the cumulative event frequencies. Thus the first entry in Table 7 states that the mean interval between earthquakes with in-tensity V11I of creater, in Southern New England is about 180 years. Table 7. Mean recurrence times'(in years) i... u , n .u.a. w 's, n... a,.. a a, 4r.. c t afi sed ,, l 'sw. W rsr.h t f. 8.tuf en ..it .fi epg4 q r. N. ort f ord waii a. . 3 iso (poi a.. tyu> w qwo> rno emi isoo (yooi
. v u i.in a. s. .r ao tuoi a q:co, im q w) isso spoon sono giocos e
is a.a.t.o too t.mo: 1.oo c oop :).m (poooi noo (.iooot soco (pooon R 7.b7 f two groop) Sooo q.'ous) iguuo (pM) loot * (poop) MOoo (61 e It is interesting to compare these mean recurrence times with the times since the last large earthquaken in the area (Table 2). The last earthquake listed with intensity VIII occurred in 1791, just 180 years , ago. Clearly, regardless of the method used to calculate future proba- , ! bilities, another earthquake of this site may be expected in,the near ' n o m m.w-. -m.q. *e -' d wa . upe . eene p m. p9. g %, g gw. p, s w. . e - w pay yim supunP p e4 . ,< y ,
4 I a 102 EARTHQUAXE NOTES future. The subject of the determination of earthquake risk from these data will be taken us in a later paper. It is worth cephasizing that the mean . recurrence times and recur-rence relations were calcul.it ed without using the large events (I > VIII) in Table 2. They are therefore independent of any errors in the intensity estimates for these large events. c' CONCLUSIONS The principal conclusions of this study say be summarized as follows:
- 1. The data in the Smith catalogs are consistent with a "b-value" of 0.95 (+0.05), applicable both to the Southern New England area as a whole, 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 The mean re-with intensity VIII or greater in the fairly near future.
currence time for events of this size is about 180 years, while the last event of this size occurred just 180 years ago.
- 4. Of the total activity of Southern New England, approximately one half is concentrated in the Boston-Southern New Hampshire region (areas A and B in Fig. 2). The remainder is scattered 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 severity of the Charleston, South Carolina, earthquake of 1886 (magnitude about 7.5, intensity about X) probably occur in Southern Nes England with a mean recurrence time of several thousand of years. There is no historical evidence to suggest when the last event of this size occurred.
- 6. Most of the large earthquakes in this area occurred during the 18th century. It is not clear if that century was unusually active, or 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 It year is 1800, and therefore do not include chis earlier high activity.
therefore possible that we have underestimated the activity of the area.
- REFERENCES Evernden,J. F., 1970, Study of regional seismicity and associated problems, Bull. Seism. Soc. Am., 60, 393-446.
Cutenberg, B., and Richter, C. F.,1956. Earthquake magnitude, in-tensity, energy, and acceleration, Bull. Seism. Soc. Am., 46, 105-145, e Isacks, B., and Oliver, J.,1964, Seismic waves with f requencies f rom 1 to 100 cycles per second recorded in a deep mine in Northern New Jersey, Bull. Seism. Soc._Am., Sji,1941-1979. Linehan, D., S. J., 1970, Geological and scismological factors in-fluencinr, the as8essment of a seismic threat to nuclear reactors, in Scismie Desien for Nuclear Power plants. R. J. flansen, Ed., The M.I.T. Press, 69-90. e iM-. . .- m i- %3. e
. . . . . . . --. -~.e . . . . . .
I 1
)
EARTIIQUAKE NOTES, VOL. XLIV NOS. 3-4 JULY-DECEMBER, 1973 10J , Smith, W. E. T., 1962, Earthquakes of eastern Canada and adjacent areas, 1534-1927, Publ. Dom. Obs. Ottawa, g , 271-301. Smith, W. E. T., 1966. E4rthquakes of eastern Canada and adjacent areas, 1928-1959 Publ. Dom. Obs. Ottawa, 32,, 87-121. I J e 4 i g-a em 9
=
ee ee e l O T 4 8 9 I , 4 . . , - , - - - - , , . , , . .g_g , - -, .- g m -
Exhibit 2 k j l t 1 T - -- * ,,-
"n 4. ,, N N .
- p.
Bulleun of the Semmological Society of Amenca. Vol 69. No. 3, pp. 757-#72. June 1979 A COMPARISON OF THE SEISMICITY OF THREE REGIONS OF THE EASTERN U.S.' BY MICHAEL A. CHINNERY ABSTRACT - Frequency-intensity data frorn the Southeastem U.S., Central Mississippi Valley, and Southern New England are compared. They are all quite parallel to one another and consistent with a slope of about 0.57. There is no evidence for the existence of upper bounds to maximum epicentral intensity in these data sets. Linear extrapolation of the frequency-intenrity data to intensities of X leads to expected probabilities for the occurrence of large earthquakes. The largest events which have occurred in these three regions are consistent with those probabilities. INTRODUCTION Recently there have been rather detailed analyses of the seismicity of three sections of the Central and Eastern U.S. Bollinger (1973) has described an extensive set of data for the Southeastern U.S., which includes the seismically active zones of Maryland, Virginia, West Virginia, North and South Carolina, Georgia, Alabama, and Tennessee, for the period 1754 to 1970. Nuttii (1974) has listed the known events in the central Mississippi Valley seismic region for the period 1833 to 1972. And Chinnery and Rodgers (1973) have analyzed the data of Smith (1962,1966) for the Southern New England region for the period 1534 to 1959. The purpose of this paper I 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., maximum epicentral intensity) diagrams. In what follows we use this type of plot, 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 selection of the boundaries of the area to be studied is much akin to the problem of the defmition of a tectonic province (which is required, for example, by the Nuclear Regulatory Comnussion Rules and Regulations, Part 100,* Appendix A). For the moment, we shall make the following assumptions: First, we assume that all subregions within a given region have a linear frequency intensity relation of the fonn log N, - a - bl where N, is the cumulative number of events in the ith subregion with intensities greater than or equal to I, and a, is a parameter describing the level of seismic activity of the ith subregion. We assume that the slope b is common to all subregions. Second, we assume that the maximum possible intenzity in each subregion, if one exists which is lower than the nominal maximum of XII, is larger than the largest 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
- The news and conclusions contained in this document are those of the contractor and should not be interpreted as necessanly representing the official policies, either expremed or implied, of the United States Government.
757
.y,,egy,, y y _ , , '9'" *W
- MNN ae, Mew MMd"==-sme s 9
s 43 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. In practice, of course, scatter in the dm often makes such a test inconclusive. However, a substantial breakdown of any si 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 diffacent data sets. As we examme and compare the seismicity of the three areas under consideration, we shall look for information related to these assump-tions. Perhaps the most important question which we shall address is as follows Each of these areas has had one moderately large earthquake in its recorded history (the 1755 Cape Anne, 1811-1812 New Madrid, and 1886 Charleston events). Me these large events consistent with the record of smaller earthquakes that have occurred more recently? Clearly, 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 eventa, which is particularly important in the assessment of the potential seismic hazard to critical 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 importance (Shakal and Toksoz, 1977). It is very difficult to document this nonstationarity within time periods of 100
- to 150 years, because of the small number of events concerned.
Tuz DATA l Southeastern U.S. Bollinger (1973) describes the seismicity of four seismic zones 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 seismic zone and the South Carolina-Georgia seismic zone. The combined area of these two zones is given by Bollinger to be 307,000 km2 . Since we would like to exclude the 1886 Charleston earthquake from consideration, we have analyzed events during the period 1900 to 1969. Even this period is probably too long for the adequate recording of intensity III 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 diagrcm is to fit the 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 similar regions, as we shall see below. The occurrence of three intensity VIII events during this 70-year period seems high, and in fact one of them has been shown to be an explosion (G. A. Bollinger, personal communication). Certainly a line ruch as the dashed line in Figure 2, which has the equation log N, - 2.88 - 0.55I (2) cannot be ruled out. The slope of 0.55 in this equation is very close to the slope 0.56 6 I
" " a o _
n .. ., ., .. .. n. : " . :sa.s. " -: -" ". -. l f - T '
'NO.VA. ! M D', i Y @n E^~~ **
A
~
n SEISMIC : ZONE ~ #8r L
&g gf ^ / .
Md O
.i:::: - _.3
~
l' - - g I . ., s-A ;
- x l W.vo
.. i x .7 . o 4
- o. _ - ,. 5 t -
-p- -. o _ l ._. .g e.; ,q --
U CENiRAL VA-
,, SElsMIC ZONE SO. dPPALCH[AN*. ZONE * . .x - ~h~ ..... *B$$ .
~~ ^
.(
a SEISMIC '{ "t; s, N . .
- 3. g i
- Y ...T m 38 9e Tenn,
- t i' yi .
is
;:f"p 3
.m
/ J TA
'k..,i",
s; - - o ,o N!C.*
. o _ kl - ~'
E l
*1
! /pt:.
_ g :- .:.
- Ih . .. im :
i h * ' '
- - 38
. mA g .= . n . . . gf.
.; . .1.. O 4 -
- teceso g 4.....
......; .. :O: g:
.. F. A IU !i Felt Reports O y j p . . ! ' Of.
.:w. figis.. l.g . #s..
Die!:'
. . - a.gJ ,#[- - --
j l g:A. j%oit - si-iii
,y _ y ,
e o S:!: -
.....w
- ^ - cz :: i 2
,, . . ;: pgij:::: . c.:
( M. ..,:j: . T .- ' ' W Vi Vil O
- g
"'s
._. ,, o
, pic
\
, [ g:sn @
v'"-" i 2 qV::: ,
% - i c- --
x
$ X u
m,
~SO. CARD' LIN A- - Two a m=*a or more $
3-j _ GEORGIA .oo .. ,. m
- o
{- o
. SEISMIC .ZONu m .__
-4 z
- s. % .
. . . . . . p
,i W {m
. ,o
.. p.. >
., .i, .. .. .. ..
,, ,. ,, e an
! Fm.1. Seismicity in tlie Southeastern U.S. for the perimi 1770 to 19%. Iteprmluted, with permission, from liollinger (i973).
- I
- t. ,
i
?
k t i._._..
l . I r ' l l
.
- e. . ==.
760 Mic,HAEL A. CHINNERY z 0.08 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 Mississippi 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 Nuttii to be 250,000 km 2. Since he lists few events before 1840, we have restricted ourselves to the period 1840 to 1969. Table 2 lists the events during this period as a function of intensity. As TABLE 1 EVENTS IN SOUTHERw APPALACHIAN AND SOUTH CAROLIN A CEORGIA 8Et3MIC ZONES latemty Period .No of Events III 1930-1969 10 IV 1900-1969 49 V 1900.-1969 46 VI 1900-1969 17 VII 1900-1969 5 VIII 1900-1969 3 os
\ SOUTH CAROLINA-GEORGE ANO
\ SOUTH APPALACHtAN 0 ~~ \ SE!SMIC ZONES 1900 -1969 a-
-o5 -
6 c. z" -i o
\
s
! \
- , ==i
' x 3, ,; ,j i I , ($ s ,-4,'o ARK. l aq,= ' =
iTENN,' t. . hl_ u m i . ,l ,
. D' .
w ......j.... A i Mo e I" *
, l / l MISS. l i
' .y i I
- 8. '. e s L .Q .* J6*
So* 87 Fla. 3. Epicenters in the central Mississippi Valley region. for the period 1833 to 1972. Reproduced. with permission. from Nuttii(1974). . s TABLE 2 EVENTS IN CENTRAL MISS(SSIPPI VALLEY tatensety Pennel Na. of Event.s II 1930 1969 22.5
!!! 1900 1969 94 5 IV 1870 1969 143.5 V 1870 1969 63.0 l VI 1840 1969 31.5 VII 1840 1969 10.5 VIII 1640 1969 1.0 IX 1840 1969 1.0 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 2 (Southern New England) and 27,000 km (Boston.New Hampshire zone). Since we wish to exclude the 1755 Cape Anne earthquake from the data set, events have been 4
,[ . !
p I
.' )- : " ,. . * ~
^
4.
, y r .7 - _;..
y _ 762 MICHAEL A. CHINNERY accumulated in both the Southern New England region and the Boston-New Hampshire zone for the period 1800 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 reasonable at to the data. In the case of the Boston New Hampshire zone, however, the number of events 1.0 ! MIS $1SSIP91 VALLEY 1840 -1969 o5 - I o - k W a: - 0. 5 - 2" Q g -10 - a
~
~ Lo% NC ' ~
' I
-2D -
II IIt 2 22 I NTENSITY Fio. 4. Cumulative frequency. intensity plot for the data in Table 2. i becomes low enough that it becomes difficult to formulate a linear fit with any certainty. A straight line through the upper four data points has a shallow slope (about 0.50), which is significantly different from the other areas studied, and which leads to high estimates of risk for large 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 timat s for risk at high intensity levels. l i
,i
a , I SEISMICITY COMPARISON-THREE REGIONS OF THE E, ASTERN U.S. 763 h k- 0 ohb.
)
O / , h O o -g $o C)@* O 0, O o r-\ O p g s!ws sQ O ' o 9g 8 e o O O' '9)
',' g
* . o) o M AGNITUDE INT E N SITY g Q
\
7 {gf7 , -Q - s
, o Q\ oCf)%l \ \ =
=
s s
..s %
O ,
*{ n__j_
)[
,* O o 4 g h ap ,. -
a :
,f e4ovina nceaoo
~f~ :
5e -
\. o
- o...
O cg 3 s p r -]{ l
-4 O O i
, ,e c 5 og\ 0 =--2 n __.
p,g . 'M -2 t e EPtCE NTRE (Uncerto'alees a 2 20') w" O .* [-
- O E piCE NT R E (uncertainties a 2 201 0 8Fivt EPICENTRES
- G SEISW0 GRAPH ST A TI O N O So soo
\ \ SC ALE iN MILES i
Fic. 5. Epicenters in New England, from Smith (1966). The solid line outlines the region called ! Southern New England in this study. The broken line indicates the Boston-New Harnpshire zone (see ! Channery and Hodgers,1973) l COMPARISON OF FREQUENCY-INTENSITY DATA l The frequency intensity data shown in Figures 2,4,6, and 7 are shown together in Figure 3. 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 cufves. There is some scatter, but each of the curves is l
' I l - -
_a -
F - 7S4 h.'CHAEL A. CHINNERY consistent with a slope somewhe/e in the range 0.55 to 0.60, and we show a slope of 0.57 which seems to be a ressonable average. In view of the rather inferior quality of much historical intensity data, it is surprising how consistent the slopes of cumulative frequency. intensity data appear TABLE 3 EVLNTS IN SotrTurnN New Exct.AND innenmry i*erwal No of Kvens 11 1928-1959 32.5 Ill 1928 1959 26.5 IV 1900-1959 43.0 V 1860-1959 24.0 , VI '800-1959 6.0 VII 1800-1959 3.0
'
- TMM 4 t
EvzNTS rN Boston.New HAMPSHIRE ZONE
. Intenmiv Penod Na of Evenw Il 1928-1959 16.0 III 192S-1959 13.5 IV 1900-1959 17.5 I V 1860-1959 12.0 I VI 1800-1959 3.5 VII 1800-1959 3.0 l 10 SOUTHERN NEW ENGLANO 1800 1959 f
f 0.5 - l . O ~ Log N C
- 2.36 -o.59 !
h-os - 9
,w ,
9
-t o.-
i 1
-is-j 6
I .a o_ l
' 1 I i t ! t I x 2 m ut I INTENSITY
' Fic. 6. Cuinulative frequency. intensity plot for the data in Table 3. ' to be. Both Connell and Merz (1975) and Veneziano (1975) have surveyed a number of estimates of this slope, and many of these are consistent with the present data. j The mean of the 11 estimates quoted by Veneziano is 0.53, but his list contains some low values which are probably not realistic. Of particular interest are the
SEISMICITY COMPARISON-THREE REGIONS OF THE EASTERN (f.s. 765 values 0.59 for the whole U.S. (Connell and Merz,1975) and 0.54 for California (Algerrmssen,19t,3). A recent estimate for the area around the Ramapo fault in New York and New Jersey is 0.55 (.,.02 (Aggarwal and Sykes,1978). It is interesting to compare a slope of 0.57 with the value that one would predict from known magnitude-intensity relationships. A selection of these relationships have been given by Veneziano (1975), in the form
~
Jf = ai + a2I. (6) Values of the constant as have been estimated as 0.67 (Gutenberg and Richter, 1956), 0.69 (Algerrmssen,1969), and 0.60 (Chinnery and Rodgers,1973; Howell, i , o5 BOSTON-NEW HAMPSHIRE 1800 - 1959 o --
. Log Nc = 2.15 - 0.591
-os -
l W , g Q. -10 - 5
' 2"o E
_J
- 1.5 -
-to -
t I ! ! !
- :Z E R I INTENSITY Fla. 7. Cumulauve frequency-intensity plot for the data in 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-magnitude data, which is usually represented by the form log N, = a - bat (7) where the slope b often lies between 0.9 and 1.0 (see, for example, Chinnery and North,1975). Combining this expression with equation (6), with a2 = 0.60, would lead to a slope of the frequency intensity relation between 0.54 and 0.C0. Clearly the
~
i _ . . . , . . . . - . . . .
3 - , . 9 l e 766 MICHAEL A. CHINNERY so, - [ o Mrs$rSSIPPi VAdEY I 6 SOUTMEASTERN U S. l e SOUTHERN NEW EELANO os- l a sosTON- NEW MAMPSMIRE j l o ..
\
\
% \
# \
$ ~ SLOPE O 57 z" \
, -io- T
\
\
- _,5- \
\
to- \
! f f ! ! I f f I a a m ga x NTEN Sl1 Y F10. 8. Companaon of the frequency-intensity data from Figures 2,4, and ".
o ulS$$$1 ppt VAUEY t5 a SOUTHE.iSTERN U S.
-% e SOUTHERN NEW EMLANO a BOSTON - NEW HAMPSHIRE to -
a l t
~,
I 25 y som s z" 9 F i a s s-l i l so-l t i I
'SF , ,
1 INTENSTY Fic. 9. The same data used m Figure 8, but normalized for the areas of the vanous zones. i I i
. . -- _ . . . _ _ - . - - . - . . . . _ - - _ _ . _ . . . _ . . . . . .,._.._...-m_.,---,_.. . . . . - _ . - , _ , _ _ , , . . . - - - - , . . , _ . - - - - , - . ,
sE!SMICITY COMPARISON-THREE REGIONS OF THE EASTERN U.S. 767 0.57 value shown in Figure 8 is eminently reasonable and con'sistent with other information. The similarity between the four sets of data shown in Figure 8 can be further emphasized by normalizing for the areas of the seismic regions. After this normali. zation, Figure 9, the recurrence curves are found to lie almost on top of one another 2 (we have chosen to normalize to 1.000 km , but this choice is completely arbitrary). The apparent simdarity in seismic activity per unit area is entirely fortuitous, and is simply due to the particular regions chosen for each study. The true 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 reason why an event the size of the Charleston earthquake could not occur in the Boston-New Hampshire zone? It is interesting to search these data sets for evidence that there may be an upper bound intensity in some of these areas. Cornell and Merz (1975), for example, have proposed a frequency-intensity curve for a site in the Boston 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 anywhere within the Boston-New Hampshire rene, the present data show no indications 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 su.h an upper bound. RANDOMNESS OF THE CATA1.OGs Before attempting to calculate the risk of large events in the three areas under consideration, we should briefly address the nature of the statistical model to be used. It is usual to assume that catalogs such as these are random, i.e., described by the simple Poissonian distribution. This problem has received ample treatment in the literature (see, for example, Lomnitz,1966). In some cases the Poisson distribution has been shown to be a good description for large events, Epstein and Lomnitz (1966), and Gardner and Knopoff (1974) have shown that the Southern California catalog, with aftershocks carefully removed, is Poissonian. Other studies have indicated departures from Poisson statistics (e.g., Aki,1956, Knopoff,1964; Shlien and Toksoz,.1970). However, these cepartures 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 least one event is given by P(T) = 1 - e-r% (8) Here To is the mean return period for events in the sample. If the time between events in the sample is the variable t, them the frequency , distnbution of t is given by l 1 - y r. F(t) e gg) ! To i l l
'1
( . ,, _ . _ >~ -
F *
~-
1 I 768 M4CHAEL A. CHINNERY lt 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 I
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 prunarily to the presence of aftershocks in the catalog. A similar plot for Southern New England data is showm in Figure 11.
! Data from the Soittheastern U.S. were not available in a form that would permit a L
')i 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 particulsr 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 deterministic meaning. , These figures are a reminder that the mean return period is entirely a statistical so MIS $1SSIPPI vat. LIV
-1900 -1972 t$ s 84 EVENTS WITH I t,Z RETURN PERICO oT
- 0.87 YEARS to T,
15 r* 4.". S
=
l t
- f"i n o 1 2 3 4 S e
. INTERocCURRENCE TIME ( years )
Fic.10. Interoccurrence times using Nuttli's (1974) data for the central Mississippi Valley. The exponential curve would be expected for a Poisson distribution. quantity, and that its only real meaning is as one of the parameters describing the probability distribution that corresponds to the catalog under consideration. THE PROBABIUTY OF LARGE EVENTS 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 periods 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 relations fitted to the frequency-intensity data, extrapolate them to larger intensities, and make estimates of the mean return periods of these larger intensities. We then use equation (8) to estimate the probability that at least one of these larger events will occur in any 200-year period, and specificaJy relate this to the 200-year period ending at the present time (a 300-year period was chosen for New England, since the largest event occurred in the 1700's).
- . . - , .m., ., - - .
1 m .
I 1 1 i ! l l l
. 1 SEISMICITY COMPARISON-THREE RECIONS OF THE EASTERN U.s. 769 I The results are shown in tabular form in Table 5. We do not pretend that these -
numbers are very accurate. In fact, because of the subjectivity that has to be used in obtaining the linear relations [ equations (1) to (5)], 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, rather than quantitative. The results for the individual areas are discussed below. 10 SOUTHERN NEW ENGLAND 1860 -1959 32 EVENTS WITH I 2 3T
, _, RETURN PERIOD To = 3.13 YEARS b6 - o 5 -
8 4 A,* M o i 2 - -- - k -- I ri i t i 1 0 5 10 15 20 INTEROCCURRENCE TIME ( years ) Fic.11. Interoccurrence times for Southern New England from the data of Smith (1962,1966). TABLE 5 Paon4stury or Lance EVENTS IN Fova RectoNs OF THE EASTERN U.S. in Peshehiirty of at Lesar one Event g g Heturn Perwal evesco T W Win m l'M T la fase tests Prewn T Yeare svttt atx ax avill atx cx Southeastern U.S.,1900- 1 23 68 195 200 99 95 64 1969 2 33 117 417 200 99 82 38 Minnsaippi Valley.184N 3 43 151 537 200 99 73 31 1969 J Southern New England. 4 29 891 3467 300 73 29 8 1800-1959 Boston-New Hampshire. 5 371 1445 , 5623 300 55 19 5 180A1959 I e The earthquake catalog for the Southeastern U.S. described by Bollinger (1973) is approximately 200 years long. Table 5 shows that, on the basis of the most recent 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 at least one earthquake of intensity X or greater will occur in a 200 year period. We conclude, therefore, that the Charleston earthquake of 1836 tintensity X, Bollinger. 197D is entirely consistent with the 1900 to 1969 data. I I a I l, _ . _ . . _ ,
I l 770 MICHAEL A. CHINNERY Without any question the largest earthquakes during the past 200 years in the central Mississippi Valley were the 1811 to 1812 New Madrid events. Nuttli (1973) lists the maximum observed intensity during this sequence as X to XI, at New Madrid, Missouri; Gupta and Nutdi (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 during a 200-year period as being about one-third. The New Madrid events were therefore reasonably consistent with the data for 1840 to 1969. Ifit 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 probability 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 1800 to 1959, Smith lists 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 Southern 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 History of the United States (NOAA publication 41 1, 1973) lists this event as intensity VIII. Other unpublished studies have deduced intensities close to VII. Whichever is correct, it cannot be said that this event is inconsistent with the l subsequent seismic record. An equally important result for the Southern New England region is that the i probability 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 consistent with the 1800 to 1959 data Notice, too, that the return period for intensity VIII is 229 years, which is consistent with the absence of such an event during the period 1800 to 1959. CoNcLttston We can make several conclypions from this study
- 1. The four frequency intensity plots that we have considered show a remarkable uniformity. All show a pronounced linearity, and have slopes which are consistent with a value of about 0.57. This,in turn, corresponds to a magnitude b value in the range 0.9 to 1.0. This uniformity, and the fact that 0.57 is very close to slopes observed in other areas of both Eastern and Western U.S., suggests that frequency.
intensity data can usefully be applied in seismic risk analyss. 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 a of those in current use.
- 2. The uniformity of the shape of the frequency intensity relation over regions ranging from the Boston New Hampshire zone and the Ramapo fault zone ( Aggarwal ,
and Sykes,1978) to the whole of the continental U.S. suggests that the problem of l 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. l
.~ .. . ._ ._ ... . . _ . . .
e l
r sEssMICITY COMPARISON-THREE REGIONS OF THE EA, STERN U.S. 771
- 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 themselves to suggest that the three large events that we ha e considered are the largest that could occur in these regions. Similarly, there are no statistical arguments that a very large event could not occur in other areas (such 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 ths possibility of events with intensity X or more anywhere in the Eastern U.S. This topic will be discussed more fully elsewhere.
- 4. The validity of linear extrapolation of the f%quency intensity data has been tested by predicting the probability of occun rice of large earthquakes in the historical record, and comparing this probabili* with the known occurrence oflarge earthquakes in each of the three areas. The ('.tarieston and Cape Anne earthquakes are both consistent with more recent data from small events (calculated probabilities of these events are 50 per cent ore more). The New Madrid 8equence is only slightly anomalous. The chance that such an event would occur during the past 200 years is about 30 per cent, but the chance that it would occur in a 300-year record approaches 50 percent. Thus, it appears that li.n ear extrapolation of frequency-intensity data to intensities of IX and X is a valid procedure in these areas.
ACKNOWLEDGM ENT Thia research was supported by the Nuclear Regulatory Commusion. The author appreciates the helpful commenta on this paper received from O. W. Nuttli and G. A. Bollinger. REFERENCES t Aggarwal. Y. P. and L R. Sykes (1978h Earthquakes, faults, and nuclear power planta in Southern New York and Northern New Jersey, Science 200. 425-429. Aki. K. (1956). Some problems in statistical seismology, Zuan 8,205-228 Algerminen, S. T. (1969). Seismic Risk Studies in the United States. Proc. World Conf. Earthquake Enr. 4th, Santiago. Bolhnger, G. A. (1973L Seismicity of the Southeastern United States. Bull Scum. Soc. Am. 63, 1785-1808 Bolhnger, G. A. (1977). Reinterpretation of the intensity data for the 1886 Charleston, South Carolina, engthquake, in Studies Related to the Charleston, South Carolma, Earthquake of 1866-A Prelim. anary Report. U.S. Geol Survey Profess. Paper 1028,17 32. j Chinnery, M. A. and R. G. North (1975). The frequency of very large earthquakes Science 190, 1197-1198. ! Channery, M. A. and D. A. Rodgers (1973L Eanhquake statistics in Southern New England, Earthquake Notes 44,89-103. l i Cornell, C. A. and H. A. Men (1975). Seismic nsk analysis of Boston, d Struct. Div. ASCE 101, no. ST10, l 2027-2443. ( Epstein, B. and C.14mnitz (1966). A model for the occurrence of large earthquakes, Nature 211,954- ! 956. i Gardner J. K. end L Knopoff (1974L Is the sequence of earthquakes in Southern Cahfornia, with aftershocks removed. Poisonian?, Bull Senm. Soc. Am. 64, 1363-1367. l' Gupta. I. N, and O W. Nuttii (1976h spatial attenuation ofintensities for central U.S. earthquakes, Bull Setem. Soc. Am. 66,74L751-Gutenberg. B. and C. F. Richter (1956h Earthquake magnitude, intensttv and acceleration. Bull. Seism. Soc. Am. 46,105-145. Howell B. F., Jr. e lD73L Earthquake hazard in the Eastern United States, Earth Mmeral Scs. 42,41-45. Knopoff, L (1964L The statistics of earthquakes in Southern Califorma. Bull. Seism. Soc. A m. 54,1871-1873. Lomrutz C. (1966L Statuncal prediction of ear *hquakes. Rev. Crophys. 4, 377-393. Nuttk O. W. t1973L The Mtumappi Valley ear;hquakes of 1811 and 1812: intensities, ground motion and magnitudes. Bull Seum. Soc. Am. 63,227-246
f . 772 MICHAEL A. CHINNERY Nuttli, O. W. (1974). Magnitude recurrence relation for central Misatsaippi Valley earthquakes. Bull. Seism. Soc. Am. 64, 1189-1207. Shakai, A. F. and M. N. Tokson (1977). Earthquake hazard in New England, Science 194, 171-173. Shhen, S. and M. N. Toksos (1970L A clustenng model for earthquake occurrences, Bull Seism. Soc. Am. 60,1765-1787. Snuth. W. E. T. (1962). Earthquakes of Eastern Canada and adjacent areas, 1534-1927, PubL Dom. Obs-Ottawa 26,271-301. Smith, W. E. T. (1966). Earthquakes of Eastern Canada and adjacent anas, 1923-1959, PubL Dom. Obs. Ottawa 32,87-121. Venezumo, D. (1975). Probabihetic and Statistical Models for Seismic Risk Analysis, M.I.T. Dept. of Civil Eng., Publication R75-34. APPt.IED Szl8 Mot,0GY GnocP LINcot.N LABORATony, M.I.T. 42 CratztoN STazzt CAMSRIDGz, MAa8ACHUSETTs 02142 Manuscript received October 17,1978 e l I I
- .- ... . ... . . . - ., .,.n..- I ~
1
,-,p- -,-..w----, v-.-,ny,,-r, .- - ...,-.-, --,, . , >-w , , . -,v -,,-,.v.-., - -., , , - , - ,-m, - . .e.---- , - . ,,.m- w-. -,e------a - - , - -n-w---
W 9 e Exhibit 3 i
+
4 i l l l i I i
)
i
= i l
l j _.- c--.
.w . - mv
. .n u::;.:f.m,.mm:.:m..
mwy;;;m . w mw.,. w
~+,m., p..
m wg. w a , n. . .w. w.y M . am w w w. .s m.. m- w.
- w,.ex x m.y w w wam w . , w..
.=.e 4-
,., . %v; . .y.. ,
. J . .,. ; , .
- , # , .: . r.
0 , ~ o V.s-s Q , :l -- .fg . s.~
- . r. ?)4. . y ?.' ' ' & .
. ..r.
m...g,A. #w.. ;g? : .m-.y.! 1
.. 9 l , - n*. ~ w . .
- 'fs #l. s -
. * . . U. . '-.
- . :+*. > ..; ^
.,. W .
4.*; *. Yt.ed
~ ~ .m s.
q % . ; *. .*'s.2, ,g' ,*Q. ;. : ;.[2,1 .s.- y. ' *+ . , _yA.. .,e m'.',1, g e.,3 . .s. y #(.i' ~ A.c,
,,.. S: *
;~g g.;* me;W/=.$:g i
' " gy. $;,.. . . . , ', ' .
3.,M.,G,. A' =: '. ,. ;r 3 p..g g. . 4Q;;. s -. .w s r*%
, M.~O6
~ + - ...<
. , ,, ; -. r- g ,.
M, u . w x., r, , wM s %. 6.%g, .e...q. , ,. w . n., ,,%. n.,.x.,n w w. :.a &' s ... w.w pfWs . hhhxy.,w.s&.b.h. hN,yN.>yN.w. % >2 %v.. . . : n .i ,:.. . m Np : . m.mM. N.$4Ihd. N.., M
' .N ;. : l tW.@$$ $$%y,2@,e.; w
.;,n.
.4 . ,. m e.
...+? r . ~.,. . . . , . ~ . v.~, . n,. n. .. . . . :.., ; .
.g ..
f_ My ..~..;Mec., W 4.,,x.y{jy.s.kk
%g,+98.m.nnRes@e@y.<@. m mn -
ny.w. .m. :~
&gw ykmf
~.
a.a s2hBW s
.-e, f,w, if A,w'X&wlf,1)$m,$1W:)w.&$h&m.$w!$hf:.[ml+ud m u. .mmmmm .. f
- wn/Q{A & . . .. s
&. ,. .m .. z%.t.m c . .. in ym 2 4 w e :; p p : w : 4
% em ? m.tyg g%Q;4rm h:%!:EkS. m& WM.M M % @r s M @p.?l$ @p.s m% 4 9%yM Q.g 1,, N.;, wm..-. . gMMW - -
n%,.w, m s%%mmass%a., w nnuaW l port
. ump.a. e M
. *am$m R%f R e
- w *q' .e1@
c-om n%, W. 3 , a. ;pm v n_.a 9
- 78.C*&?N %: ' ' $_s.w._f.MW4WMA.n.i_m@,e&&'igQ
- w. . . , ... w. .us. . .
- 5,M,C. % ;5.'i
.m g ,NQ{,$d hs... b .s[w .
~
M;* maw %.W.Wc.mtM% ummge:uq%.h.a 4 a my ..nm
' . ..M&v%4,Mm%: : ...
M M Te q q th421 hw
. w. : . Qk:r; x . ... z. t.% w y.s u p m. # m s. % e !.MMS we. g -0. W 4. g q. _ g. gw:p n ~s m.,4:n g.%,%yn
- u. - 2 g. 9 ., ik Q. , a%g &
q 3 x. ~ n. m%%c.q.< yes,tigationsM 3ecSeismolo~gicalg;
~ r? m r.I n. m w ~ g-n .'I - -
nputy g+. g g g . M M M I, j %g g g" W % . M ! N d i $ M fei f 6 N [ M @ N $ 1 E @ N .$ F $ ew w MgweWBX England w $ n..wqswo. 4 m e g @tg re&Q Wmg&wwReactorsvin.-New h h !eww;h.wp,g k h .,$
- w. ..a_ .w. m+. .,.w.-
., ..s .. ,, m..
. z. ,
...e,.
, . . wwm+..#,.em' s.-.,
%. v .. . . . . .. n. . . m. . ., . . ,. e Tr. .~..,..%~.
,.'t,. g. . . .. . .
e
.m. c. s < s .
%. ,s,m.,,,s..
,s,.,
. . .s .
- w. . e ,....r... .
- w. . n.s..,..
.w,,.,.. ....,,3.e ..,
. . u . ,3. w g ,, ,.%
4,, ;,,
?hf; b- k$ h
...~ .
.1 -
h . .,b b' [ ho Yb, k A, $h . k*AA W r e - }Q[&lll$&. d- PA. Pre
$,$...G $'h~ w h7g. , $g@f,. ?SkYWWOi&?.5W&S38.#= M%v@M$; i%n$& . .:
~. . m.n -
.. n .
2-7 4$$e=.M h mn= .w ' v q :m m.Lvn -inCO_n La~3 oratory ~ l\ . . ..
,. / , i ,3 . ,
3.
~,
. MASSACHUSETTS INSTITUTE OF TECHNOLOGY I
Lte cTos. mssAcm;stTTs 021:3 ' . N'
.,t7.
-e s
. /
.w 8.-
,g
% 97
..-W g 9-L 5, .
-- . . *5 ,PP' '# * .
' / "s. 8 .
- a fc
.s ,' . ;< e ,..o g ,
e,.,.6' .c .;;j , -? s .w. ./- . < 2,. .3,.w>
-7.g ,.a'.%;.
- 1 e p
' X.. .3:n? y*
.. f, .
,.4
;y - . /. w; < .
, ,, .- . .. .. ~ s
, , , .. 3
,,,.*'e3" rd' d * * ., . *
~
b )* l', - i .b a ** #' O' ' U. 8 "M
. ., , t
%.;t.Lq'...&..m.y'3 D .
-l .ppN,M w .. . .'rQ . [ '1. ,, .;;; .&
_ %[ ' d?.g 3. :. f . . *} y*L, .. ,
, . f * "_*R. Q.;~ ..;'..~. H<'\,.>l,;. '
w:= - p~ . = . m .3 N.,r
,- , M'T l' w .( 3 ~ q. . s
- d. . . . .*,,,,;,,,.3...m.
,..J,~
' , } g. h :... . ..
-. . . . .. . .t.
@^ #
[' .;h ~ )-.g -di} l ji, , hs...+ ', ':i.@, . ' *i c ,~
....'4. h
- .y. h P,%, Q ", . " , .,, rc,;'*?
n = < .t.. i s ,- ,.* . ,
. a .g^.., . . . , .* +-(
- b. .*g.aw.s..a;.e . :k. c., .* ., ...,.S..9
. . ,d ?,p.;4,cc. 4 ( . .y. . - O .;4 , . .. . . . . + r ;.g.
- .y , i,s. ,
. . v
.4"
.n 1', am.ue* * *< . 4?
. . , .y g J.,;-.wn./ b. r .w' ' a.w . :,. ,
.,o.n,.y; , ys C. .*
r.' W. we - y (. i3
, .s t. .e.. .a ,'+/ J y Q ,zm .- 4;p=* ,s.,
3.g .,,a.., @ .,f m*' N me- ,.
- .m, '. - . e w,,'.,p.s u~m-v-
.,m. w . m .
.s .-4 . m*,
n, <W" ,' l P,r t ,, y j > u . ~s,, .~ ~. u.
,,R. ,s.g .
s r,,
.c.::w%q . ,e,.
; w- Q. . get*t'ny:MWyM ~v..,.PM.."j %.fn(JM,,e e .6 s g.#*-m.- ,3~*ruLp~v.:. &n , . ,
w v m.s,: e.. < h%n.? Q M e:g'. W ?'l% . &..,s.
~ ..
,9,, w . x m :. .
. w.n .nJFin5 M.%.i.f MWn,M. e.. '
% ?. R, m.
- 3
, w. ~.. w - m e. c . w. , v:,w .a . ., . m.
Massachusetts Institute of Technology Lincoln Lnboratory AN INVESTIGATION OF MAXIMUM POSSIBLE EARTHQUAKES Annual Report Project
Title:
Investigations of the Seismologicalinput to the Safety Design of Nuclear Power Reactors in New England. NRC Contract: N RC-04-77-011 Principal Investigator: Michael A. Chinnery, Group Leader Applied Seistnology Group Lincoln Laboratory, MIT 42 Carleton Street Cambridge, MA 02142 Period of Contract: 1 January 1977 - 31 December 1977 I l 15 August 1978 l l l l l I . - . - .
Abstract This report describes research c'arried out under NRC Contract NRC- ) Oh-77-019 during the period 1 January 1977 to 31 December 1977 A detailed study of available scientific literature concerning the estimation of maximum possible earthquakes shows that all available methods are empirical and lack a sound physical basis. Evidence that even.the empirical methods'are valid is very weak, primarily.because of the short j 1ength of the earthquake record in most areas. An attempt to use global earthquake catalogs. to examine the regional variation of maximum possible - earthquakes is unsuccessful. It is demonstrated that saturation'of the ! magnitude scale and biases introduced by instrumental clipping combine to make m values for large earthquakes very unreliable, and to obscure b the presence or absence of maximum possible earthquakes. A progress report on a study of New England crust and upper mantle structure is
! included.
i e i t t l iii > t. L ,_ .-, _ - - _ _ , . . , , _ . - . . , - . . , . . _ , _ . - , - . , , , , _ . . _ _ _ . , _ . _ _ _ . _ . , _ . . - . , _ . _ . . . _
, r L
. Table of Contents Abstract ' iii
- Introduction 1 1.. Maximum Possible Earthquakes: Current Status 2 1.1 Introduction 2
~
1.2 Definitions 3-1.3 Approaches to the Problem 5 1.h Physical Arguments 7 1.5 Arguments Using Earthquake Statistics 10 1.6 Use of the Level of Seismic Activity 12 17 Pattern Recognition Approaches 16 1.8 'Other Studies 18 1.9 Discussion and Conclusions 19
- 2. Analysis of Global Catalogs 23 2.1 Characteristics of Global Catalogs 23 2.2 Earthquake Statistics 2h 2.3 Saturation of the Magnitude Scale 28 -
2.h The ISC Catalog 32 2.5 Events in the Aleutians-Kuriles Region h0 2.6 Interpretation 49 2.7 Discussion 57 ( l 2.8 Conclusions 60 References 61 Appendix: Progress Report: New England Crust and Upper i Mantle Structure 68 iv l D - ~ = - , .. - - - - -m m_ u.m ..mmmm, , 3, , , , , ,
1 Introduction This report describes research carried out under URC Contract NRC-I Oh'-77-019 during the period 1 January,1977 to 31 December,1977. The major effort during this period consisted of two studies aimed at evalt. ' ing P the possibility of estimating the maximum possible earthquake that mi m -g
. be expected within a given region.
The first study consisted of a review and' assessment of available scientific literature on this topic. Since much of the research in this area has been carried out in the Soviet Union, this review provides a reasonably comprehensive set of references, and a discussion of the various approaches which have been tried.
! The second study was an attempt to look for evidence of upper !
I bounds to earthquake size within global body wave magnitude catalogs, and in particular in the ISC catalog. This study soon turned into an atte=pt to understand the sources of bias in the magnitudes listed in 1 this catalog, since until these are understood it is impossible to search for maxinun possible events. It transpires that these biases, together with saturation of the g eale, make g catalogs essentially useless for this type of study. A third area of research, into the crust and upper mantle structure of New England, got underway during the period covered by this report, and a progress report is included in the Appendix. i I . i e 'k k " w yw- M e g-g+'we re-y,r-g-4vw-9 y- g- -m-ge-_ v p p ms -p-vp-g--- * - --_e-- gpe- s >y gw k g- bw-r g
2
- 1. MAXIMUM POSSIBLE EARTHQUAKES: CURRDIT STATUS 1.1 Introduction We vould like to know whether or not there is a limit or " upper '
bound" to the size cf earthquakes for a variety of reasons. First, earthquake size is usually intended to be a measure of energy release. However, energy usually varies strongly with size. For exmaple, the standard relation between magnitude M and energy E (in ergs) is log E = a, + b,M (1.1). Bath (1966) reviews several estimates for the constants ao and bg, and shows that b g appears to lie in the range 1.4 to 2.0. Since the number N of earthquakes is usually described by the relation log N = a - bM (1.2) where b is about'l (see, for example, Richter 1958), the total seismic energy release is dominated by the largest events. We shall have reason to question both equaticus 1.1 and 1.2 later in this report, but the. conclusion appears to remain valid. Analysis of the energy budget of the earth requires knowledge of the rate of occurrence and energy release in the largest events that occur. Second, Brune (1968) has-shown how the relative slip of tectonic plates can be estimated from earthquake size, and showed that the total slip is dominated by the largest events that occur. The fundamental question of hov much tectonic motion is released in seismic slip (Davien l l , and Brune, 1971) can only be answered clearly once ve understand these large events. And, thirdly, the esti=ation of maximum earthquake size is impor-tant in the estimation of seismic risk. The possibility that large events may occur, even infrequently, in an area can lead to a seismic e __ _a-_ _ w -> b.__
. . . - . _-. . - -, . - - . -. . . _ . . . ~ . . .
5 3 .
. 1 hazard that is' unacceptable for certain critical facilities such as- ,
nuclear power plants. The NRC Rules and Regulations, Part 100,. Appendix A, set out the seismic safety standards for these structures, and define the Safe Shutdown Earthquake to be based on an' evaluation of the " maximum i earthquake potential" of an area (Hofmann, 197h). .The purpose of the ; present study is to assess our ability to estimate this quantity. We can usefully divide the overall problem into tvo parts. First, i what is the evidence that earthquakes considered as~a global phenomenon have a maximum possible size? And second, how'does this. maximum possible size vary from region to region? The first question ought to be much 1 simpler to answer than the second, and it is logical to examine it-first. However, as we shall see, it is difficult to give convincing ansvers to either of these questions. 1.2 Definitions "! There are two important definitions that ve.must explore before've continue. The first is the definition of " maximum", and the'second is the definition of "siza". The term " maximum" is not, unfprtunately, always used with the same I b meaning. One definition is the obvious one, which refers to the largest i possible event that can occur given the physical conditions of the
+
source area. A second definition, sometimes used, includes the concept
?
l l of probability. A certain probability level may be accepted as being
" negligible", according to engineering design standards or other arguments, >
and the " maximum" earthquake defined as one which vill occur with this probability level (or less) during the projected lifetime of a structure. I
! These two definitions cre very different, and it is essential that ;
l ,g they be clearly distinguished from one another. We shall use the terminology I L - )
h M for the "true" maximum possible magnitude (E for the max 1=um o possible energy, etc), and M for the =cgnitude that occurs with probability P, which defines the accepted probability of "negligibility". I 1 As we shall see in the next section, very different methods must be used in the estimation of M and M . The definition of earthquake " size" is even more difficult. There are a large number of quantities which attempt to measure this size. A partial list includes: a) Body wave magnitude (mb) b) Surface wave magnitude (M,) c) 100 second period magnitude d) seismic moment (M ) e) radiated seismic energy f) elastic potential ener6y release g) maximum epicentral intensity (I) h) maximum epicentral acceleration
- 1) local magnitude (g)
The basic problems here are not only to decide which of these measures of size are the most appropriate fo$ a given situation, but to recognize that the relationships between these measures are in general poorly understood and in so=e cases demonstrably very non-linear. In parti-cular, some of these quantities have built-in upper bounds which can obscure the search for a fundamental upper limit to earthquake size. We shall examine this problem in more detail in section 2. An additional complication, which arises in the literature very frequently, is that the term magnitude is so of ten used without proper definition. All practical measures of magnitude are restricted to some
- . - - . - - - - - -- .. .-. ~. ..-
F -
/
5 l limited portion of the seismic spectrum, and are closely tied to the i method of measurement employed. There is so much variability in both of these factors that the term. magnitude alone is almost meaningloss, particularly when the characteristics of large earthquakes are concerned. ! Quite often, in reference to the local seismicity of area, the term
, magnitude refers to local magnitude b(. Of all measures of magnitude this is one of the hardest to quantity. It was introduced originally by Richter, and designed for local shocks in California. Its definition is very arbitrary, and refers to the logarithm of the maximum recorded trace amplitude of.a specific instrument (Wood-Anderson seismograph) at a specific distance (100 km). Because the instrument vill record a vide range of frequencies in the short period band, and because there is no seismic phase identification, the significance of the maximum trace
~
amplitude is not clear. For small earthquakes, the maximum trace amplitude i vill often refer to body wave arrivals at short distances. For istge earthquakes, the maximum trace amplitude vill usually be associa+.ed with fundamental mode or higher mode (L phase) surface waves. The principal usefulness of Mg is, of course, that it is a measure of ground motion in the near field at a range of frequencies that are relevant to engineering considerations. Improvements in the estimation
. of M (Kana=cri and Jennings,1978) may lead to a more consistent scale, but its relation to far field magnitude determinations is still unclear.
1.3 Accroaches to the problem The nanber of papers in the literature that attempt to get to the heart of the problem of the estimation of the maximum possible earth-a quake is quite small. The majority of these are the work of scientists in the USSR, where there has been a long term interest in this topic. 1 1 I
. . ~ . . - - . . __ . . . . .
6 Unfortunately some of these papers 'are hard to obtain and difficult to l read. . A number of approaches _to the problem have been proposed (see, for. example, dhenkova and Karnik, 197h). First, there are a number of broad arguments that attempt to limit the upper size of earthquakes on the i basis of physical principles, including fault gaometry and slip, and the strength of earth materials. Generally speaking, these arguments make a convincing case in favor of a gicbal upper bound, but give little indication where this might be. A second approach uses earthquake statistics, either in the form of frequency-magnitude data or modelled by the theory of extremes. These two analytical techniques generally lead to sbnilar ] results, but both turn out to be severely limited by the definitions of magnitude used. A third approach, which seems very logical yet which lacks any convincing physical basis, attempts to relate the size of the maximum possible earthquake to the level of seismic activity in a region. It would be very nice if such a relationship were to exist, but there is no clear evidence that it does. More recent approaches have tended to focus on information from non-seismic sources, such as geological and
~
_ geomorphological data. Some of these approaches are statistical, using pattern recognition techniques. Others are more deterministic, and attempt to link long term geological fault movement to short term earthquake slip. In virtually all of these approaches one problem predominates. The record of earthquakes is relatively short in most parts of the world. Data before about 1900 are generally qualitative and hard to interpret. Adequate seismic networks have only been available since the early 1960's, and (as we shall see in section 2) there are still proble=s in , I
,- ,.m , we r -isw e , ,, ., ,we ,w.y- . ,e- m ,-e+- e w , .- w e --- r e -i , +,w-,. ,--n ev
. - . -~ .-
9 7 defining the size of large earthquakes. It therefore becomes very difficult.to establish empirical data for maximum possible earthquakes in specific. regions, since these largest events may have return periods of 1000 years or more. Without these empirical estimates, it is virtually impossible to examine the validity of many proposed approaches.
. 1.h Physical Arguments There seems to. be universal argreement that any measure of size of an earthquake must have an upper bound. This argument is often intuitive, but it can be refined to some extent. Certainly equations 1.1 and'1.2 cannot both be valid for indefinitely large M, since this would. imply an infinite release of seism'c ienergy per unit time-(Newmark and Rosenblueth, 1971). However, both of these equations are poorly defined at large magnitudes, so the argument is not too helpful.
Intuition is often carried into the discussion of regional upper r: bounds. Newmark and Rosenblueth (1971) remark that earthquakes with M > 9 in the continents and M > 7 under the deep oceans are unlikely, though they a6 nit there is no real basis for these estimates. In fact, if M is surface wave magnitude M,, we shall see that M probably does not exceed about 8.6 anywhere, but this is an artifact of the magnitude scale and not a true upper bound (section 2.3). Earthquakes of M, > 7 'i' have been observed several times on the mid-ocean ridges, where the o activity is lov. .j-Sometimes intuition is quantified by the use of Bayesian statistics ; Connell and Merz (197k,1975) propose an upper bound to earthquake f epicentral intensities in the Boston area on the basis of a presumption I i
! that such an upper bound exists, and " conversations with seismologists".
The resulting seismicity curve is used to estimate seismic risk in this
5 8 area (see also Esteva, 1969; Veneziano, 1975). It seems likely that this study reflects a general belief that areas of lov seismicity shoulit have low upper bounds to earthquake size (see section 1.6). It is possible to go somewhat beyond intuition. Tsuboi (1956) has proposed an upper bound to earthquake energy. He first relates earthquake energy to the volume V the strained region around the source, then assumes that the strain is uniform throughout this volume, and then uses field evidence for the maximum strain which the earth's crust can withstand (about 10 ). Then, if V is limited by the thickness of the crust, an , upper bound to energy of about 5 x 10 ergs is obtained. It is hard to assess the validity of the assu:nptions used in obtaining this result. A very similar approach has been given by Shebalin (1970), though it is less convincing. He quotes linear relations between earthquake magnitude and both mean length of focus and vertical extent of focus, from an earlier paper (Shebalin, 1971). He then uses limitations on both length acd depth to set an upper bound to magnitude. The validity of his starting relations is very much open to question. Similar procedures have been out? sed by Hof" Ann (197b), Who describes how magnitude fault-length relationships (e.g. Bonilla and and Buchanan, j l 1970) may be used to assign maximum magnitudes. Obviously this type of I approach presupposes that we can clearly define the location and length , 4 of all active faults in an area, that breakage beyond the present fault i i length is impossible, and that the magnitude-fault length relation is single valued (this is equivalent to proposing that all earthquakes have the sa=e stress drop). Each of these ast.umptions is difficult to justify. i l Shenkova and Karnik (197h) raise the possibility that the rate of I strain accumulation =ay set limits on the maximun energy released in an g ti- e-w s y g g - -mir- y y e-w-rg- <y--g--ty-y ,- gg-gry v- w- gr y-pv-- -m- +- e-*e- v--m-4 tw-== w---e- m sein- -
--f
( . i a J
.9 l
earthquake. They indicate, for exanple, that if' upper and lower bounds can'be placed on a Benioff strain release graph,'the maximum possible earthquake vill be specified. This approach is meaningless unless'the record of earthquakes already contains at least one mav4== possible - 1 event. These studies are typical of those attempting to use physical , arguments. The strength of rock, under various physical conditions,'is : not well known. However, we know even less about the limitations.ontthe size of the zone of slip, and it is this variable which probably limits t the usefulness of physical arguments. The largest known fault area is probably the 1960 Chile earthquake, which was about 1000 km long and perhaps 200 km vide on a shallow dipping fault plane (Kanamori and Cipar,197h). There do not seem to be any convincing arguments why fault breaks could not be larger than this on occasion. Could the -l. entire Aleutian are system break at once, for example? The effect of strength of rock is related to stress drop. The basic problem can then be formulated as follows: Seismic moment M,is defined by - - Fi g
= pLWD (1.3)
- where p is the rigidity, L is the length (long horizontal dimension),
- W is the width (shorter vertical or devn dip dimension), and D is the
+
average fault offset. 3 . The stress drop Ao can be written Aa = n (1.h) where n is a geonetrical factor which typically ranges from 0.25 (for long strike' slip faults) to 0.75 (for long dip slip faults), as is shown by Chinnery (1967).
.t *W- =*(g- 6 i- -Se t- og g y-v. i, v+q yvy yc-w y 4- e e v a-g m . .- y- a e:r 3-,wu-q.er+r-- y v u we o. - t 91 mes re:i .e .c g*n.e- w..w ,. ae ww.w .b-si'ew rr- *'y +4 t FF" e-*W T--*, g e y m.e -a a i epeg , ,t.--gr,-v g ,w m. y + n . eeuw. q w-=.*% _ +y +,iw-1 4
l l 10 So ve may generally write M O 2LW Ac (1.5) If stress drops are roughly the same.(about 50 bars) for all earthquakes,1 as has been suggested (Kanamori and Anderson,1975), then limitations to seismic moment M depend only on limitations to the dimensions of the fault area. However, questions about the constancy of Ao remain. Some studies appear to indicate local stress drops as high as several kilobars (Archam-beau,'1978). In the eastern US, the occurrence of modcrate sized earthquakes in the lower crust with no surface expression cf movement vould appear to require rather smal] fault dimensions and correspondingly large stress drops. To take an example, if a fault area 20 x 20 km vere , possible in an area of stress concentration in the Eastern US, with a , stresa drop of one kilobar, equation 1.5 gives a seismic moment of over 10 dyne-em (equivalent to an M, of over T.5, see Figure h). This is probably larger than any earthquakes so far observed in this area. We conclude, then, that while physical arguments support the idea that there must be an upper bound to earthquake size, and suggest that there may be a substantial regional variation of this upper bound, we cannot yet constrain the appropriate parameters enough to estimate the sizes of these upper bounds.
- 1. '5 Arguments Using Earthauake Ste,tistics A variety of authors have attempted to use the statistical char-4 acteristics of the earthquake record to estimate maximum possible earth-quakes. It is not at all clear that existing earthquake catalogs are good enough for this type of study. Certainly, in the example discussed in detail in section 2 of this report, it is clear that problems of 9
m m m --,-,---o--- , vr--wv- ,- , , . - - * ,w- ,e mr - prw-g-a ~ s a ,- -e
11 I saturation of the magnitude scale and individual station detection- + completely obscure the presence or absence of upper bounds. There are two possible approaches to the analysis of earthquake ' catalogs. The first involves the use of the frequency-cagnitude curve, which is discussed extensively in section 2. j!'he other is based on ;
, t Gumbel's'(1958) Theory of F.xtreces. Gumbel described three asymptotic distributions which may be used to model the distribution of largest events occurring in a sequence of equal time periods through the earthquake ,
record. The Type I asy= ptotic distribution of largest values corresponds , to a linear frequency-cagnitude relation, with no upper bound. The Type II asymptotic distribution includes the case where large events are less frequent than vould be expected on the basis of smaller events, i.e. a
'non-linear frequency-magnitude curve. '"he Type III asy= ptotic distribution specifically includes an upper bound. Algebraic details can be found, 6
for example, in Yegulalp and Kuc (197h). Applications of the Type I distribution generally acco=plish no more than the use of linear frequency magnitude statistics, and no upper bound is included. Papers using this distribution include Epstein and Lomnitz (1966), Gayskiy and Katok (1965), Milne and Davenpert (1968), Connell (1968), Karnit and F.ubnerova (1968, 1970), Yegulalp and Kuo (197h), Shenkova and Karnik (197k) and Shakal and Tokso: (1977). ':"nough l some of these papers =ention maximu= cagnitude earthquakes, it is clear e that what is discussed is the quality !? sax, the magnitude which has a probability of occurrence (during sc e .ixed period) that is less than P. Studies that atte=pt to use the ?ype III asymptotic distribution are potentially more interesting. These include F'ei-shar and Lin 3- r- , - - - - < yy,. -w- - - p - - =w g,m< -s,y-.
12 (1973), and 'legulalp and Kuo (197h). The first of these studies does not define the magnitude used, while the second is based on Gutenberg and Richter's (195h) data. They can both be shown to be formally equivalent to trying to fit the frequency-magnitude curve with a truncated distribution (Cosentino ett al., 1976, 1977). We note that Knopoff and Kagan (1977) have argued that frequency-magnitude statiatics are to be preferred over extremal statistics since the first uses all of the available data. To anticipate section 2, there is no doubt that saturation of the It is interesting to note that most M, scale begias in the range T-7 5 of the estimates of M from these studies are greater than M, = T.5, and the vast majority are greater than M, = 8.0. As long as saturation of the magnitude scale is not considered, there is no way that the results can be unambiguously interpreted as indicating the presence of an upper bound with regional variations. 1.6 Use of the Level of Seismic Activity Ferhaps the most persistent attempts to study the nature of er.rth-quake upper bounds hsve been made in the USSR by Riznichenko and his co-verkers, beginning with Ri nichenko.(1962, 196ha, 196Lb). Many associated references are listed by Riznichenko and Bagdasarova (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 . release E = 10 joules, he discusses the problem in terms of E 3, ar.d K,. He uses an inplied relationship between energy and the observed quantity, magnitude, of the form log E = a + bM (1.6) The particular values of a and b used are not quoted (and are still open to question), and the particular definition of magnitude M is not given.
~
s p.,
. 13 He . recognized from the beginning that it was difficult or impossible to determine K directly from the observed earthquake catalog of an.
area. He has therefere focussed on the possibility of establishing a relationship between K and the level of seisnic acti"ity A in the frequency-energy relation
. log N = A - y(K - K )
7 (1 7) (A is therefore the activity at the reference energy level K,). He has discussed the form of the relationship A(K ) in several papers (Rizni-chenko 196ha, Rznichenko and Bagdasarova 1976 and others). Briefly, his argument is to relate the energy K of an earthquake to a volume radius R (for Central Asia he obtained R = 0.315 10L 10), to average the activity A over a circular region of radius R to obtain I, and then determine an empirical relation batveen I and K . For Central Asia he determined (Riznichenko and Bagdasarova, 1976) log'I = 2.8h + 0.21 (K=ax 15)- (1.8) while for Japan he found a better fit with log I = 2.8h + 0.39 (Kcax -
- 15) (1 9)
These equations are inter.ded to be valid for 15<K<19, or 1022<E<1026,,g,, The form of these equatiens was derived very artificially (R1:ni-l chenko, 196ha). K was simply chosen as the largest event fcr a given region (often using a short time sa=ple), and I determined for the regien. The plot of X against K had considerable scatter, and a linear relation was fitted to the largest K values (Riznichenko and ,l Zakharova,1971) . In 196h the constants estimated in equation 1.8 vere 2.80 and 0.20, so there has been little change in the relation in the
! subsequent 12 years. The difference in the slope found for Japan (0.39 instead of 0.21) is disturbing.
*y 6" *
- L i
i lh
- t k
r l I b i log N log N = A - yK ! P K K = log E max log E max Fig. 1. Riznichenko postulates a clear-cut upper bound to total earthquake energy E, and assumes a linear frequency-energy relation a for energy values belov E . l
# 'M '
*=w'+'--*= w y .~ g , g, , , _ , , _
w a w A c e 15 ) 1 Obviously, the problem in this , approach is that K , needs to be ! l determined in some regions before the general law can be established. .l l l We must allow, however, the possibility that successive application of ' the equation in various regions (e.g. Gorbunova, 1969: Drumya and Stepanenko, j l 1972) may improve the constants by an iterati,ve or " boot-strapping" l method. The logical basis for the expression 1.8 is not established. Whether or not it works in practice is less clear. The authors compare 31 large earthquakes in Japan with the predictions of equation 1 9 Twenty-one are found to be in agreement,10 are found to be larger than the predicted K , , though the authors note that uncertainties in many of the epicenters make it hard to make a firm conclusion from this result. The situation is far from satisfactory. The existence of a relation between K and A is not proven, and appears to be more of a hope than max a scientific fact. We should note, in passing, that if the maximum value is defined using a probability P (E , ), then there is a very clear relation between the maximum value and the rate of ~ seismic activity. This has been described, in a most obscure vay, by Housner (1970). His argument may be restated as follows: Let us assume a linear unbounded frequency-magnitude law of the form log N = a - bM (1.10) where N is the cumulative number of events, with magnitude > M, per unit area, during a unit time period (per year, say). Suppose that N is the number of events / year that can be considered negligible for risk purposes.
, . - - - , - e . , . _
._ _ o ., ..,.y , . ..,. , , ,y, _y , , _ . , , - - , ,
e 16 Then log N =a-b[ (1.11) For two different regions, with different a and b values, we have log Nn " *1 1 max " "2 max 2) so, b a -a [ max (2) = b 1 [ max (1) + b (1.12) 2 2 , It is reasonable to set b : b: 1, and then 2 (2) = (1) + (a 2 - "1 or 2 N [ max (2) = [ max (1) + log (1.13) N o where N is the number of events with magnitude 0, which may be taken as an indication of the level of activity. In a simple example, if area 2 has a seismicity of one-hundredth of area 1, then the [ value for area 2 vill be two units smaller than the [ for area 1. 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
- 1. Clearly the analysis really refers to our unbounded frequency-magnitude law.
In s - a g , existing literature sometimes attempts to postulate a relationship between seis=ic activity and the upper bound to earthquake size, but success in establishing the nature and even the validity of this relationship has been essentially non-existent. 1.7 Pattern Recognition Accroaches , 1 Recognizing the fundamental difficulties involved in trying to relate the size of maximum possible earthquakes to the level of seismic e-t i af Om Mi &Et m
17 activity alone there have been several attempts to include a variety of other geophysical and geological info'rmation. Ricnichenko and Dzhiblad:e (197h) have compared and correlated the estimation of K uing the level of seismic activity, the gradient of the Bouguer gravity enomaly (suggested by Tsuboi, 19ho, and Berg el g., 196L), and the ve:ocity of vertical movements determined by geodetic and geomorphological methods. The three estimates were combined together to obtain a single estimate using weights of 1.0 for the seismic data, and 0 5 for each of the other methods. The results are no more convincing than those based on seismic activity alone. This paper is notable, however, for its extensive collection of references. Shenkova and Karnik (197h) state frequency-energy data are not reliable enough for the estimation of K , and urge the inclusion of data on " environmental properties and the rate of energy accumulation" I (i.e. Benioff graphs). However they giv little indication how these pieces of information should be tied together. In view of the interest of several Russian geophysicists in pattern recognition problems (see, for example, Gelfand et al., 1976), it is not surprising that attempts have been mm.de to apply these methods tc the determination of M . This topic is addressed by Bune et al. (1975), and an application to the Carpathian region is described by Borisov and Reysner (1976). ne general idea is to look for those combinations of f *
. observable features that appear to be indicative of the observed M i
values. The features selected include such items as rates of recent I vertical motion, nearby volcanise. presence of fractures and fracture intersections, seismic actP 'ity anomaly etc. The data analysis follovs the usual procedures. Est of the features chosen were found to vary strongly with M . I
i l 18 The basic proble= of this analysis is, however, not addressed by the authors. In order to deduce the ap2ropriate relationship, values of ; known M , are needed in a substantial nunber of regions. Since these are not readily available, the authors used " estimates made by experts". This introduces such a strongly subjective element into the analysis that it must be regarded as meaningless. 1.8 Other St.udies Tvo recent studies should be mentioned, the first for completeness and the second because it has an interesting approach to the problem. Caputo (1977) has proposed a complex model which purports not only to account for the linearity of the frequency-magnitude relation, but to predict the~ maxi =um seismic magnitude and moment. The assu=ptions on which the author bases his analysis appear to be completely unreasonable, and the paper is meaningless. i Smith (1976), on the other hand, has proposed using geological data l
! to obtain a mean rate of slip for a fault zone over the past 10's of thousands of years or longer. Then, if the frequency-moment relation-ship for the area is linear, and can be defined (see Chinnery and North, 1975; Smith's argunent here is less rigorous), then there cust be an upper bound =ocent that is consistent with observed slip (Brune, 1969).
S=ith uses geological data of Hamilton (1975) to obtain these upper bound coments (which he converts tsck to upper bound magnitudes). This approach is one of the most reasenable that we have seen, but problens still remain. There are considerable difficulties in the definition of the frequency-coment relationship for a limited zone. Even if this can be esticated, however, there must still be difficulties in the interpretation of geological slip data. Slip on the San Andreas l s l l l . -- _ -. _ _ _ - . - - . _ ~ ~
19 fault syste has clearly been distributed over a rather vide zcne on a geological time scale. It is likely that irdividual faults could carry much of this slip for a period of ti=e, and then it could be transferred to other neighboring faults. Tc put this another way, Smith's (1976) approach requires that the earthquake process be stationary over the
, period of the geological data en each fault c,onsidered. This is a questionable assu=ptien for the fault :One as a whole, and may be invalid for individual faults within the systes. And, of course, there appears to be no way to apply S=ith's =ethod to regions such as the Eastern US, where geological information on fault slip is not available.
19 Discussion and cenelusions The basic problem in attempting to deter =ine the envinum possible earthquake in a region can be stated quite simply- If the earthquake record for the regit:. has a length 7 years, then evidence is available that bears on the earthquakes that have mean return periods of up to T s years, or a probability of occurrence devn to 1/T per year. This evidence is not necessarily good evidence, for the largest earthquakes in the sa:ple. The occurrence of large earthquakes appears to be descr! bed quite well by a poissen distribution (Epstein and Lc=nitz, 1966; Lennit:, 1966). The probability that at least ene event with an annual probability of 1/T vill occur within a peri:d of t years is
~ . - ,
"~
p=1-e (1.1L) So, if t = T, the probability is 635. This suggests that in more than One third ci all regions studied there is likely to be an apparent deficiency of large events. t i i 0
_ __ . -_.. . . _ . - . _ . - . _- =_
. i 20 ~!
To phrace this another way, a 100 year record of earthquakes vill i only give reliable information (at the 9,0f level) for those earthquakes with a mean return period of about h0 years or less, or an annual pro- . i bability of .025. or more. In practice, of course, the length of the - earthquake record is often considerably less than 100 years, and this applies to most of the regions of the USSR studied in the quoted literature, and to California and other active zones. Clearly, then, a 100 year-record of seismicity is only adequate for the determination of maximum possible earthquakes if the mean return periods of these earthquakes are significantly less than 50 years. This implies that the maximum possible earthquake must have occurred several times during the period of observation. In all of the literature that has been surveyed, there is no case of a specific region where a maximum possible earthquake can be clearly defined. Even when all regions are considered together in a global earthquake record, the apparent upper bound to surface wave magnitude M, can easily be accounted for on.the basis of saturation of the magnitude scale (Chinnery and North, 1975). Perhaps the most useful contribution to this area that could be made at the present time would be the clear and unambiguous demonstration of the e'xistence of an upper bound to earthquake size in just one region, anywhere on the globe. It is necessary to add, here, that we have not attempted to define the term " region". This is a thorny tc% c (see, for example, Hadley and I i Devine, 197h) which has been emphasized by the term " tectonic province" which appears in the NRC Rules and Regulations, Part 100, Appendix A. We shall not discuss it further here, except to note that given a cap of epicenters for the earthquakes in a seismic zone it is always possible to select a region that contains no large events. The validity of such a selection is very questionable. M MW, WMMMde,.*blGW---
)g 48 $ 4 t- P# 9 W Md9 4t@ 4 y* i $ ps gg q -p .g$ sh gengeq$ t4Q g g etway e g q wg$ q e g gp gggr- __
. 21 It appears, then, that existing seismic data are unable to throw any light on the questions of the existence and size of maximum possible earthquakes. In spite of the deep seated belief of many seismologists and earthquake engineers that upper bounds must exist, the only reasonable approach, given our current state of knowledge, is to assume that these
~
. upper bounds are at rather high levels in all areas.
'de are therefore forced into the classic method of simple extra-polation of linear frequency-magnitude or frequency-intensity relation-ships. This raises an additional probles which deserves discussion.
In the context of the evaluation of the seismic risk to critical structures such as nuclear power plants, ve vould like to establish a way to determine the size of the earthquake that occurs with some fixed risk probability within a given region. Following McGuire (1976) and others, we may 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 mean return period of 10,000 years. If the process is non-stationary, this statement is meaningless. However, in practice ve have very little alternative but to assume that the avail-able record of earthquakes is representative of the rates of occurrence of both small and large earthquakes in the immediate past and the immed-iate future. The problem of stationarity is not easily set aside. Evidence from very long conpilatiens of earthquakes in the Mediterranean area and China (the latter was discussed by Lee and Erillinger, 1978) show disturbing changes l.n seismicity on time-scales of a few hundred years. The seicnic record in New England shows similar changes during its 300 year length (Chinnery and Rodgers, 1973; Shakal and Cokso:, 1977). Clearly this , ,. . , . . ~ -.-m. y_., , - --. , ,-- ,. .y w, , ,. - .-+ y = . . ~ , # _ , - , , , ,
22 raia s the pessibility that large earthquakes =ay be associated with sone long term average level of seismicity which is very_different fro = the recent short record of staller events. It is important that research into the stationarity of earthquake processes in various tectonic environ-ments continue. The most prc=ising avenues for future investigations into marimum possible earthquakes vould appear to lie in three areas. First, we need more infor=ation on the nature of the strain and stress fields in seismic zones. Second, we need to improve our understanding of the ultimate strength of crustal materials in a vareity of tectonic settings. It seems likely that the true upper bound is centrolled by the size of the region of acec ulating stress, and the ability of the crustal rock to withstand that stress. Tr.irdly, the information from geological and geomorphological data on long term fault slip, where surface faulting is visible, =ust place some constraints en the largest possible earthquakes l (Smith,1976). This approach needs further developnent, though the l question of stationarity may li=it its usefulness. L 9 e l l t:
- . - - -. . ~ . _ .- .- , . _ . . . . . - . - -. - . - .
. , e t
, 23 ,
i
- 2. /JIALYSISOFGLb3ALCATALOGS
'2.1 Characteristics of Global Catalogs t
A legical place to seek for infer =ation on the existence of upptr '! bounds to earthquake size, and the variation of these upper bounds with ;
,i tectonic region, is within earthquake catalogs. There are basically two r
kinds of catalogs, those compiled for a limited region using data from a 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 thf s seems most likely to contain evidence for regional variations, if they exist. [- In order to be useful for this study, a global: catalog must have i .
?
two important characteristics. First, it cust be complete, particularly ; for large' earthquahes, and preferably for medium-sized events as well. - i Second, it must use a clearly defined measure of earthquake magnitude ; which is unifor=1y applied to all events. As we shall see, this turns , out to be a =uch more restrictive condition than it appears to be at first sight. . r Several global catalogs are available. These including events P 1 since the early 1900's include Gutenberg and Richter (1951.), Duda (1967) t and Rothe (1969). Unfortunately, the gic .a1 distribution of seismic n stations was very poor until 1960, and the se catalogs all suffer from a ' high degree of non-homogeneity. With the establish =ent of the World i Wide Standard Seis=ograph !!etwork (WSS?i) in the early 1960's, a cuch , , scre nemegeneous data set became available. Data frc= this netwerk, together with a variety of data from other stations were analyzed by two ; organisations. The U.S. Coast and Geodetic Survey, and its successors the :iational Ocean Survey and the U.S. Geological Survey, have produced l.
+
1 m,,m g-e v e,r v . ,t v e m*,e, m,-9,o-y-*,-,.-w,ev-*,,--+re-+w*e eo-,-ww---- -ev e- e' w-*w e-- rv+m=~***=-*-a-es***"*w"*"'+*-*~r**'*N*+ * ** * * - * ' * ' * * ' ' * " * " " " * "
- i 2h -l 1
a fairly rapid bulletin (the PDE, or Preliminary Determination of Epicenters) l issued on the average about 6 months after an event occurred. The International Seismological Center (ISC) has chosen to collect all the available data, including the PDE bulletin, and issue a more comprehensive catalog. Typical delays in the publication of the ISC catalog ranged l from two to three years. Both the PDE and ISC catalog began consistent routine bulletin production at the beginning of 1964, and since then have maintained the production of very uniform catalogs. Both catalogs, since 196k, have recorded a body wave magnitude ab for essentially all events. This magnitude is based on the maximum peak to peak amplitude in the first few seconds of the P-vave arrival en short period instruments (operating in a rather narrow frequency band centered at about 1 bz). Surface wave magnitudes M, (at a period of about 20 seconds) were recorded very irregularly, and only in the last year or two have attempts been made to measure M, on a routine basis. The requirement that the catalog be co=plete forces us to focus on the body wave magnitude3m . For reasons which are outlined in the aext sections, this is not desirable, but there is little that can be done about it. Attempts to relate M, to sb have shown a large scatter (see, for example, Aki, 1972). 1 In the sections that follov ve shall concentrate on the ISC catalog fer a very practical reason - it is available in detail on magnetic tape (the detailed PDE listing is not). This facilitates a variety of computer analyses of the very large amount of data concerned. l 2.2 Earthauake Statistics There are two basic representations of the statistical characteristics of an earthquake catalog. One deals with the relationship between s- e++t-*s-,- :p - w + a e e n e 9 .iy- -,,,.m.-+,mr,,,,e-iar*- -e c emww,g-e++-wme+,i ,,mmwee-*,.vm-e.-+e.., a
.=.g-ww~--%a'-WS*- 'td em w'* w- D*-=tw-w-~4"e- e w?'-'Y-W*wm+-y'**--'-'p-es'**W '
-WT-
1 25 ; earthquake frequeney.and earthquake magnitude. The other utilizes ! Gumbels (1958) theory of extremes, and is concerned only vith the largest . event within a given time period. Though these'tvo approaches appear to be very different, they give very sLnilar results when applied to the $ same data set (see, for example, Chinnery and Rodgers,~1973, and Shakal and Toksoz, 1977). Because of this, and because the frequency-magnitude - ,
. app' roach uses all of the data in a catalog, it is to be preferred. ,
t Knopoff and Kagan (1977) have specifically shown that extremal statistics are much inferior in some cases. For this reason, we shall use the frequency-magnitude approach throughout. Gutenberg and Richter (see Richter,1958) demonctrated that local earthquakes in California obeyed a frequency-magnitude relation of the , form: ;
't ' log N g =a 0
- bM .(2.1) where Ng is the number of earthquakes with magnitudes in a small range centered on M, and a and b are constants. This form of the equation is necessarily discrete (the constant a depends on the size of the magnitude intervals in which the earthquakes are accumulated). In many cases, it is more convenient to use the cumulative form:
log U **~ *
. c where, now, N e s e nun er even s v t magnitude M and greater.
This equation may be regarded as being continuous, and is more amenable to analysis. It is easy to show that if equation 2.1 is valid, then equation 2.2 is also linear and has the same 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-magnitude curve, and it must be regarded as empirical. Even cr--M.-eweg., vi.-y+r+-e-w-., .,9y q .y , w,6 ,v.-,.sg7,y,,g,1,."*** e w NF = r 49'T M T"'e t to~ se k eNwg=***W"F. T' d N U W h 4'"'t'ee*=*T*'=*f-' f*'-* t w T
- 26 I
i I i log N e linear frequer.cy-magnitude law
\
s
. M Magnitude en Fig 2: Ideal effect of an upper bound to ,
earthquake magnitude, using cumulative frequency-cagnitude statistics. l
, 1 l
1 e I
1 27 18 2-12586 l 1000- - i
~
100 r o -
. x -
s _ m c _ W W
~
10 -- o - z _ . W _ . D o _ LOG 10 N = 7.66 -0.93 M s s. , W
\.
u. w 1.o - \.\ . H
\.g J -
o _
\y 2
o -
\
1 0.1 r ~
\
k ,
\
~
\
\
1 0.01 I - I ' I i 6.0 7.0 8.0 9.0 MAGNITUDE (M3 ) , F1g. 3: rata frc Gutenbers and Richter (195k)- i e --eye- .,w- g- -wq +e---v+ g t'v 7 --~*eW--- e -e---- e y 'vw-- e r- +=wameary y w re--r-n e-a aw- +-i-w e +m e -vgr ***m--rp'-e-- m*+=i ** *1+Meaw-ww
_ - ~.- . --. - - 28 using observational data, the universality of a linear relation is not clear. Many of the reasons for this vill be discussed in the sections , that fo11ov. In an ideal world, the presence of an upper bound to earthquake magnitude vill reveal itself by a departure from linearity at the upper end. Figure 2 shows an idealised representation of this non-linearity. Unfortunately, there are two other effects that can'also lead to a curve similar to Figure 2. First, any measure of magnitude based on a limited i spectral band has a built-in saturation property. This is discussed in the next section. And second, seismic instruments frequently have a , limited dynamic range, and the magnification is often set to record medium sized earthquakes. In this case, large earthquakes vill cause the instrument to go off-scale, and a measure of magnitude is impossible. As a result, there may be a purely instrumental upper-bound to measureable magnitude for a given instrument. The effect of this on network determina-tions of event magnitude is discussed in later sections. 2.3 Saturation of the Magnitude Scale Several authors (Chinnery and North, 1975; Kanamori and Anderson, 1975, etc) have recently pointed out that because of the shape of the spectrum of the radiation emitted by an earthquake source, any measurement of magnitude based on a limited spectral band of frequency must saturate. For example, Ms is usually measured at about 20 seconds period. When the source is large enough that fracture propagation lasts for longer than 20 seconds, the amplitude of the 20 second radiation vill not change with increasing size, though its duration in general vill. An exemple of this effect was discussed by Chinnery and North (1975). Figure 3 shovs the cumulative frequency magnitude curve for
29 large events listed in the classic study of Gutenberg and Richter (195h). It appears that the listed magnitudes are very close to present day M, values (Evernden, 1970). This diagram has often been used as a basis for discussing the existence of an upper bound to earthquake magnitude (see, for exa=ple. Housner , 3 970 ) . It is, however, possible to interpret this curve in , another way. Figure k shows a compilation of recent data relating surface wave magnitude sM to t.ie seismic me=ent M . The highest two points correspond to the 1960 Chile and 196h Alaska earthquakes. Both have been extensively studied and seem reasonably reliable. The observa-tional data clearly indicate a saturation of the M, scale which seems to begin at about M,=7.5, and be complete at about M,=8.5 The solid line in Figure h is a rough form of the M,-M relation. At this point we can legitimately ask if the fall-off in Figure 3 can be wholly attributed to this saturation. We can say this much: if the data in Figure 3 are translated into a frequency-moment graph, the result is very linear (see Figure 5).
~
Kanamori and Andersen (1975) have argued that the frequency-moment graph should be linear, with a slope of 0.67, if all earthquakes have
, the same stress drop. It therefore seems reasonable to postulate that this is the case, and to conclude that the Gutenberg and Richter result
~ (Figure 3) can be explained as saturation of the M scale. s There are two important points that arise from this study. First, c:. a global scale, there is no direct evidence for an upper bound to seismic mocent, though McGarr (1976) has argued on geometrical grounds that such an upper bound must exist fairly near the highest mcment data j point on Figure 5
i
! 30
,\
!8 2-12585 31 10 i
- CHINNERY. AND NORTH 1975
! o CHEN AND MOLNAR 1977 3 10 - I 29 . - 10 _
,_. . o E o 28 e. o
, # 10 _ l E = l 3 - 8 0 H
.o
} Z 27 *
- w 10 -
o 2 . ! O t 2 . 6 26 3 10 - 25 g
...I .
- 8* .
10 - E
*s.*
24 . , , , 10 , 5.0 6.0 7.0 8.0 9.0 MAGNITUDE (M3 ) Fig, k: Cc=pli.'.atien of 87 published estimates of seistic sc=ent as a function of surface vave magnitude M s-9 e __.-mw, - . , ~w ,--+-e yw -. = p- yy e.c., -e,p g y=_ . . mai- ,,,--,.,.,.w,
, , , ... ,- g ,..ig , + ,
31 18-2-12587 1000-1 F - I 100 - m
. o r W -
A N - _m c - W W
~
10 -
>. P 1 o c l z -
w a o r - w t I W - L i- w 1.0 - F- - 4 r
=j ~_ LOG 10 N = 17.47 -0.61 LOG 3g M o .
2 3 - U e 0.1C-L
~
l F
~
l O.01! 24 25 26 27 28 29 30 31 LOG 10 (moment) Fig. 5: Frequency - :: ent Ernp ::nstru:ted frc: Figures 3 and h.
32 Second, the importance of =agnitude saturation is demonstrated. When ve come to examine global catalogs using the 1 hz m s ale, ve must b expect saturation to occur at lover magnitudes. This vill clearly make the problem of trying to estimate regional variations in maximum earth-quakes very difficult. 2.h The ISC Catalog An incremental frequency magnitude plot of data in the ISC catalog for the period 1966-70 is shown in the lefthand portion of Figure 6. Although ISC data are available for a longer period, we have chosen to limit ourselves to this 5-year span in order, as ve shall see, to compare the overall catalog with certain special stations that were only operating during this time. The resulting plot is typical of all frequency-m data currently b available (e.g. Brazee and Stover, 1969, Brazee, 1969). There is no clear linear portion to the graph, and this has led some authors to propose a non-linear relation (e.g. Shlien and Toksoz, 19T0; Mern and Cornell, 1973; Stewart, 197h). It is therefore very difficult to determine a unique b-value, though typical atte= pts to do this lead to high values of up to 1 5 or more (see Figure 6). At low magnitudes many 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 # 6.6. No events larger than 6.6 appear in the catalog during this time period. It seems reasonable to ask if these catalog characteristics are in any vny the result of the stations used in the analysis. As many as 500 or =cre stations feed data in to the ISC, many of the= very irregularly. To examine this question, we selected a subset of 28 stations which operated continuously throughout 1966-70, and which report regularly to d - As w-
33 the ISO. The statiens used are listed in Table 1. Magnitudes vere recomputed as the average of those reported by the 28 stations, and a requiremont that at least 3 of the stations must have reported the event was superi= posed. The resulting frequency-=aghitude graph is shevn in the righthand portion of Figure 6 (the solid points). A second data set ,
. was for=ed by applying the station =agnitude biases determined by North (1977) to the 28 station network. The results are shovn as open circles. ,
The 28 station network shows very sinilar characteristics to the ' catalog as a whole. In particule.r. the general curvature of the graph and the fall-off at high =agnitudes are preserved. This is convenient since it allevs us to study the 23 station network instead of the whole catalog.
There are reasons to suspect that biases may be introduced into the i netwerk =agnitudes by the precess of averaging the reported station magnitudes. This proble= will be discussed in more detail in later secticns of this report. It suggests, hcvever, that it may be worthwhile locking at the frequency-n. characteristics of the events reported by o
individual stations.
. Figure 7 shews plots of the events reported by Kevo, Finland, for ,
1966-70. On the left are counts of icg A/T values (A is the observed a=plitude of ground =ction, and T is the cbserved dominant peried}, i which are independent of scurce location. The values are converted into ) 1 statien =b by the application of a standard anplitude cistance ccrrection. This ccrrection is test kncvn in the distance range 30 to 90 -degrees, and the righthand side of Figure ~ shows events in this distance range. Similar data for Fort Meresby, : ev Guines, are shcvn in Figure 8.
,e - - - - - - , - ~ - , --w--w ,_.-.--.,,,.--,,iw- .~~..w- ,9 -9y- -..mve..ww.ww w ar+- - ip - -- -ym., .v-.-- 9
t 3h TABLE 1: 28' STATION NETWORK STATION CODE LOCATION BIAS (North, 1977) ALQ Albuquerque, N.M. - -0.20 BEA' Broken Hill, Zcabia -0.28 BM0 Blue Mtr.s., Oregon -0.29 BNS Bensberg, Germany +0.20 BUL Bulawayo, Rhodesia -0.07 CAN Canberra, Australia -0.02 .t CLK Chileka, Malawi -0.27 COL College, Alaska +0.01 COP Copenhagen, Denmark +0.36 EUR Eureka, Nevada -0.2h KEV Kevo, Finland +0.02 - KHC Czechoslovakia +0.10 KJN Kajaani, Finland +0.1h LJU Ljubljana, Yugoslavia +0.29 MBC Mould Bay, Canada +0.1h MOX Moxa, Germany +0.02 NOR Nord, Greenland -0.1h NP- Northwest Territories, Canada 0.00 NUR Nurmijarvi, Finland +0.19 FMG Port Moresby, New Guinea +0.10 PRE Pretoria, South Africa -0.07 PRU Czechoslovakia +0.0h RES Resolute, Canada +0.13 SJG San Juan, Puerto Rico +0.2h TF0 Tonto Forest, Arizona -0.32-TUC Tucson, Arizona . -0.1h UB0 Uinta Basin, Utah -0.11 WIN Windhoek, South Africa -0.09 4 l
c22-5sau 1966-70 3000
~ ALL EVENTS ' 28 STATION NETWORK oOO 3 STATION DETECTION 1000 .
r . 6...O* O SLOPE 1.49 - WITHOUT STATION i : j
~
- O BIAS ;
~
' ~
.O O o WITH STATION BIAS
. O i
' O ICO ? N :'
. E SLOPE 1.47
~
~. O i
' O' l 10 .
1
-o -
3 , O } } i , , i i i e i i i i i - i 3.5 4.0 i 1-4.5 5.0 5.5 6.0 6.5 7.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 - 7.0 4
*b b i -
I Fig. 6: Frequency magnitude data for the ISC catalog, for all listed events (left), and for a 3 selected network or. 28 stations (right). The 28 station network is listed in Table 1. i
- i 4
c22-5583 1966-70 4 1000-ALL EVENTS ; 30* < A < 90* y , y [ - ' r l ~ t' l i y i ! 10 0 - I
~
SLOPE 1.41 ~ .l SLOPE 1.45 1 i N - - i - j . . . j 10 r ,.
. 'r 1 . -
. u -
y
, -. .t 1 1 \
j 1 i .. i i i il i i i i i L . O O.5 ' l.0 1.5 2.0 2.5 3.0 4.04.5 5.0 5.5 6.0 6.5 7.0 1 i LOG A/T STATION m b 4 j Fig. 7: Frequency magnitude data for Kevo, Finland. i 4 e .
PMG ca2-ss36 1966- 70 IOOO' ALL EVENTS - 30* < A < 90*
.) _
~
/
/ %s -
~
I SLOPE 0.97 A. . SLOPE 1.03 10 0 : I / : f
/
1 i- p N - g 1 . lo r f r . \
.\
g
%\*
\
*\ *
\
~
t . .\
\ \ \
1 . i . . i i, _. . . . . . (, 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 LOG A/T STATION m b Fig. 8: Frequency magnitude data for Port Moresby, New Guinea.
-. . - - - _. . . ~ - .. . _ . . .- .
t 38
'We have compiled similar plots for all of the stations in the 28 station network. A vide variety of behavior is seen. If attempts are made to fit the frequency g plots with a straight line, slopes are found to lie anywhere within -he range 0 9 to 1.5 Figures 7 and 8 show
~
clearly the differences that are observed. There are two possible interpretations of these data. If the diff rences in b-value are real, this could indicate an important regional variation in seismicity characteristics (clearly PMG and KEV sample different portions of global seismicity). 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 made for the U.S. VEIA observatories. These are BMO (Blue Mountains, Oregon), UB0 (Uinta Basin, Utah), TF0 (Tonto Forest, Arizona) and WMO (Wichita Mountains, Oklahoma). The four plots are superimposed in Figure 9 Each station has been adjusted horizontally according to the station biases of North (1977), and small vertical adjustments have been made to improve coincidence, recognizing that there are small differences in the seismicity sampled by each station. Again, only events in the distance range 30 to 90 are included. Remarkably, these data are all consistent with a seismicity curve that is linear, with a slope of about 0 9, up to m =5.8, and then the b curve bends downwards and approaches the vertical in the range m =T.0 b to 7.5 This relation, indicated as a solid line on Figure 9, is remarkably similar to the Gutenberg-Richter M3 curve (Figure 3) in shape. However, it differs dramatically from those observed by normal stations. Notice, for example, that these observatories record many events in the range l b=6.7 to T.2, whereas ncce are listed in the ISO catalog. , I _ . _ , _ . _ _ - . . . . _ _ . . _ . , . . _ ~ . _ . _ _ _ _ . _ . _ . . . _ _ _ . _ _ . . - . . _ , . . _ . - . _ .
I 39 I 5000 l18 2135251) I Oog x 0 ,a 100C - A o e VELA ARRAYS
- xA ADJUSTED FOR STATION BIAS
, AND SEISMICITY LEVEL o ,' 1966 TO 1970
. A O '
u 0 A 4 10 0 _
~
N - \ SLOPE 0.93
- \
O 6 N N
- BMO 3
\
~
O UBO #*aO\ \ i x 0 \ x TFO O e 10 - a \ A WMO * \ ' AA % \ O XX g O g en N I I i I I f , i I I 35 40 45 SO $$ 60 65 70 7.5 m b
, Fig. 9: Frequenc'/-station .g. picts for four U. S.
VELA obser ratories. s--,ryr +es' r--vv*- y t--w--t,-rwr --y r e- e ve rw wg -wr- wyw- g e w,- w-c <g q--+w 3 reg-y-+w- * - - - . -T--+wwwwWte+ywW,,--S-*we-r' --pw o mW- wr roesy5 pg+w* * =r-+--of- 't yr
I I f h0 There are a number of important diff,erences between the VELA arrays and the average analog seisnic station. The operators of the VELA arrays were highly trained specialists, who nade an unusual attenpt to measure magnitudes carefully and censistently. More i=portant, each of the arrays was equipped with a lov gain channel, which gave the arrays a cuch larger dynamic range than the average station. These points strongly
- suggest that the VELA data may be more reliable than regular station t
reports. An additional suggestion that this is the case is obtained , from the Large Aperture Seis=ic Array (LASA) in Billings, Montana. Figure 10 shove data from this array for a completely different time period (1971). The seismicity curve shown in Tigure 9 is an excellent fit to this data set (in Figure 10 this seismicity curve has been adjusted I vertically for a best fit). In order to investigate this probles in more detail, it would clearly be advantageous to limit the geographical regien within which the events are located. In this case we may expect a well defined seismicity curve, and ve 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 The analysis of the previous section was repeated for events in the Aleutian-Kurile Island area (defined by longitudes 135 E to lho V, and latitudes 30 -90 ). The i=portant seismicity of this area lies within 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-cagnitude data do not disagree strongly 'rith the seis=icity curve shown, which is that shown in Figure 9 adjusted vertically fcr a best fit. Upon closer examination, it transpires that the catalog for _ = - - . _ . , . . . . _ . _
41 c22-5623 1000 --
~
LASA BULLETIN
, 1971 ASSUMED BIAS = - 0.25 100 N -
I 10 - _ c 1 1 I I I_ l I 3.5 4.0 4.5 5.0 S.5 6.0 6.5 7.O M AGN!TUDE (mb) 1 ! Fig. 10: Frequency-magnitude data for the Large Aperture Seismic Array (IASA) in P.ontana for the year 1971. The solid line is the seismicity curve shewn in Figure 9
~
I
~ --
I i k2 C22-5621 1000-. _- e ALEUT AN-KURIL EVENTS -
- 1966-70 e
ALL ISC e e 100 -
, e, e
N _ e o 10 -- e a e
- e o 1 1 I I I I .I 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 MAGNITUDE ( mb )
Fig. 11: Frequency-nagnitude data for all events in the Aleutian-Kuril area listed in the ISC catalog, l 1966-70. j 1 1
1 h3 4 i this area is heavily biased by the reports from the VELA observatories, particularly for lov and moderate events. ' The situation is clarified in Figure 12, which shows the data for a , twenty-five station network.(this is the same network as that listed in i i Table 1, with the VELA sites BMO, TF0 and UB0 removed). As before, ' o. three station detection is required before an event is included. Nov , the shape of the network curve is clearly very different from the seismicity curve of Figure 9. In fact, it is very difficult to locate the seismicity curve in any "best fit" position by vertical movement. On the other hand, data from the VELA arrays for this area show ! P excellent agreement with the global seismicity curve, as shown in Figure- i
- 13. Notice again that the VELA arrays record many events with magnitudes '
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 location of the' stations used, since the*e are 6 North American stations included in the 25 station network. f i We can accentuate the problem further by considering only stations ' I in Europe. Figure lh shows the same data for a 10 station European " network which is listed in Table 2. 3 The addition of the biases of t North (1977) do not change the disagreement in shape with the VELA stations, but they do reduce many of the network magnitudes. This results from the generally positive bias of European stations (Table 2). ] If the postulated seismicity curve (Figures 9 and 13) is real, there are clearly problems with the magnitudes reported by the individual stations in the network. As an exa=ple, Figure 15 shows the observations of Aleutian-Kurile events by station KEV (Kevo, Finland), which was discussed earlier (Figure 7). Either the reported magnitudes are subject
----wy,,-,-swmvg- vp ' n~ < '# *w-97 4
I hk c22-562T 1000 - ALEUTIAN -KURll EVENTS 1966-70 TWENTY-FIVE STATION
. ., NETWORK
. 100 - j Z * . N _ , 1 10 - .
~
1 I I I I I f 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 (MAGNITUDE (mb I Fig. 12: Frequency-magnitude data for a 25 station netverk (the stations listed in Table 1, with .BMO, TFO and UB0 omitted) .
h5 C22-562h 1000 - 3 ALEUTI AN-KURIL EVENTS
. 1966-70 OA 3 . UBO o o TFO
. A o A BMO 100 --
0 A
.o N -
A
~
A o o o 10 *^
- p.
3 0 0 _ . 40 0 A O Cl 1 I l l l l ob.o .. 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 MAGNITUDE ( mb) Fig. 13: Frequency-magnitude data from 3 VELA arrays for Aleutian-Kuril events. The solid curve is the same as that in Figure 9, ad. justed vertically for a best fit. i
k6 l c22-5625 1000- _ ALEUTIANS-KURIL EVENTS 1966-70
]
_ TEN STATION EUROPEAN NET
- WITHOUT BIAS e.
O O
- O. o WITH BIAS 100 -
o
- O.
O.
- e
_ O N - O. . e 10 - E o O e
- 0 0
- O
- . O e
- Oe
] l c. . I I I I n -
3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 M AGN ITUDE ( mb I l Fig. 1h: Frequency-magnitude data for a 1C station European network. The stations used are listed in Table 2.
~
kT TABLE 2: 10 STATION EUROPEAN NE7a'ORK STATIONCbDE LOCATION BIAS (North, 1977) BUS Bensberg, Gemany +0.20 C0P Copenhegen, Denmark +0.36 IGN Kevo Finland +0.02 KHC CJechoslovakia '
+0.10 KJN Kajaani, Finland +0.1h LJU Ljubljana, Yugoslavia +0.29 MOX Moxa, East Gemany . +0.02 NUF Numijarvi, Finland +0.19 FRU Czechoslovakia +0.0h STU Stuttgart, Gemany +0.29 9
e
h8
. c22-5626 1000_
= ALEUTI AN -KURll EVENTS
_ KEV _ 1966-70 100 - __ e _ e N - e 10 - z
- eo
_ e
- o
_ e e
- ee se 1 I . I I I i 1.
3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 MAGNITUDE ( mb I Fig. 15: Frequency-cagnitude data for events in the Aleutian-Kuril area, as o'eserved at Kevo, Finland. The solid curve is the same as those in Figures ll-lh. E - - . . ___
h9 I l to strong biases, or the station is failing to report many large events. For the reasons discussed in the next section, the latter explanation ~ 1 I seems most likely. 2.6 Interpretation At this point we are faced with two possibilities. Either the U.S. VELA arrays (and perhaps LASA, too) have a poorly calibrated low rain
. channel, which leads to the systematic overestimation of the magnitudes of large events, or the magnitudes of these large events is systematically underestimated by the global network of analog seismic stations. We have been unable to find any independent evidence for the first of these alternatives, and it must be considered unlikely. It is possible, however, to suggest an explanation for the second of these alternatives, based on the dynamic raage of typical analog stations, and the process
' of averaging which is used to obtain a network magr.itude. Any seismic station can be described by a detection probability curve. The general form of this curve, and the parameters necesse.ry to define it, are shown in Figure 16. For our present purposes, since ve are examining an earthquake catalog, we should regard this as the curve describing the probability that the station vill report an amplitude of an earthquake to the analysis center (e.g. the ISC). If, for exa=ple, the station does not operate for a portion of a given time period, the maximum probability P vill be less than 1.0. 3 The probability curve falls off at both lov magnitudes (where the signal is not nessureable) and at high cast.itudes (when the instrument is off-scale). T.7e 50% detection levels can conveniently be used to define the dynsmic range of a given station. Nctice that in practice the location of these points vill depend to scce degree on the diligence w -,w , - - y
50 18-2-13521 H J __ p ___ R SATURATION
$ DETECTION OR m
o N . CLIPPING m CL Z )$ -
)$ -
o H l . [ M l l-. W I o b Gs STATION m b d , STATION DETECTION PARAMETERS G H D d 7d SPREAD OF DETECTION CURVE Gs 50% SATURATION THRESHOLD' 70 SPREAD OF SATURATION CURVE , B STATION. MAGNITUDE BIAS P PROBABILITY OF REPORTING R 1 Fig. 16: Form of the Detection Probability Curve l for a seismic station. . _ _ . , . _ . . _ ,, . . . _ . . . . . _ _ . _ . _ . , _ . ~ _ . . , . . _ . _ _ , . . . _
51-l of the operator. This is an additional complication which is hard to i model; though it may be one of the most important effects in determining ; the dynamic range for amplitude reporting. Amplitudes are generally measured with a rule on the seismogram, which is. traced by a beam of light on photographic paper. The smallest . ~ amplitude measurable depends on the line thickness, which is typically- . about 1 mm. One would expect a=plitudes of a few millimeters to be easily measurable. With larger events, however, problems arise. Most ; i operators record the amplitude, zero to peak, of the first- sving of the > trace. When this intersects the edge of the paper, most operators' vill not report an amplitude. Also, when the trace amplitude becomes more ; than a few em, the ability of an operator to locate the tip of the peak ' (or trough) vill depend on the quality of the photographic recording, which is-usually quite variable. And very large events, even if they do l not go off-scale, are usually difficult to =easure. On purely geometrical grounds, one vould expect the dynamic range of amplitude reporting to be between 2 and 3 orders of magnitude (i.e. t between 2 and 3bm units). As ve shall see, however, it seems'to be between 1 and 2 orders of magnitude in practice, and "cceplete" recordings , of amplitudes (the flat part of the detection probability curve) is usually limited to less than 1 order of magnitude (sometimes much less). Where the station probability curve vill sit on an absolute a b scale vill depend on the station magnification and the statien bias 3 (the latter are seldo= more than a few tenths of a magnitude unit: see Table 1). The station detection probability curve has then to be considered in the light of scattering processes in the earth. These are illustrated v , , . - w-. ..m . -
52 in Figure 17 Because of scattering, an event of =agnitude m vill lead to a distribution of observed magnitudes at a network of stations. This distribution is often roughly nor=al with standard deviation about 0 3 g units (Von Seggern, 1973), and its mean (in the absence of station bias) vill be an estimate of m. However, when the cagnitude of the event approaches either the detection threshold or the clipping threshold of the stations, the distribution becomes skewed. Effects near the detection threshold have been discussed by Ringdahl (1976) and by Christoffersson et, t g (1975). Those stations where scatter-ing produces a lov amplitude vill not report, whereas those where a large a=plitud'e occurs vill report. This leads to a net positive bias when the station reports are averaged to produce a network magnitude. Methods can be devised for including the fact that scce stations did not report an event (the maximus likelihood =ethod) but these methods are cu=bersone, and require a detailed knowledge of the detection probability curves. It does not appear possible to apply the:a 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. When a large event occurs, those stations where scattering produces a large amplitude vill usually not report, while those stations that receive a lov a=plitude vill report. The result is a negative bias to the netwerk =agnitudes reported for large events. This negative bias vill be quite substantial, up to 0.5 or 1 magnitude unit, and can adequately account for the difference between the VII.A seismicity curve and the ISC catalog seismicity curve. I N _
- .- . .. . . ~ - - _ -
53. Number of Observations i I l Approximately normal l
/ o - 0.3 mb Units I
l i m Coserved Station Magnitude Fig. 17: The effect of scattering on observed amplitudes. An event of magnitude m vill 1ead to a distribution of nagnitudes at
, a network of stations. This distribution is usually appr ximately normal. I l
l l 9 _ - - - ~
Sh We can illustrate our argument by usf.ng data from a single station. Figure-18 shows the data for EUR (Eureka, Nevada). The left hand portion-of this figure shows a conventional interpretation of the reporting characteristics of thic station. An arbitrary straight line is fitted to the data, and detection and clipping thresholds (indicated by arrows) are determined at m b
=h.5 and 6.3 respectively. In the right hand portion of the figure, the VELA seismicity curve is used (EUR is quite close to the observatory UBO). In this interpretation the station fails to report many events for m greater than 5 5 b
The thresholds are now h.3 and 6.1, and " complete" reporting is limited to the range k.T to 5 5 A similar interpretation for staton KE/ using Figure 15 suggest that this station carries out " complete" reporting over an even an11er range, perhaps as little as 0.3 m units (fr s 5.2 to 5.5). b A different representation of the same phenomenon for station EUR is shown in Figure 19 Here, for each interval of 0.1 mb. units of UB0 reported magnitudes, we have averaged the difference in reported magnitude between EUR and UB0 for events in the ISC catalog during the period
~
1966-70. The theoretical interpretation of such a data set has been discussed in detail by Chinnery and Lacoss (1976). If the detection probability curve for EUR vere hori: ental (Figure 16) then this plot should be horizontal too. The presence of a detection threshold shows as pronounced positive biases as lov magnitudes. There is a hint of a flat portion of the curve in the vicinity of 5.0-5.5, and then the data continue beco=ing more negative. This must be interpreted as being due to a clipping threshold. In general terms, Figure 19 is entirely consis-tent with the right hand preferred interpretation of Figure 18.
c22-5587 EUR 1966-70 1000 : 1 :
/ EN l.
[4N 100 r a : *
*\
_. _. .g
.\
~ ~
. .. .\- -
\
10 - g i-
\ :
\
i , u\
~
\
\
4 \ 7 y. 1 i i i i i li i i . . . _k. ' 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 7.0 STATION m STATION m b b Fig. 18: Two interpretations of the reporting frequency of station EUR (Eureka, Nevada). The right hand interpretation is preferred.
56 c22-5h61 e 0.5 - EUR - UBO o 1966-70 co . o
,e W
O D e b z . O e 2 e i O.O - e E
- B e ee a e w
O
- o ee g_
e e Z e O -
<C
- 2 - 0.5 -
*e
! I I I 1 3.5 4.0 4.5 5.0 5.5 6.0
, MAGNITUDE (UBO)
Fig. 19: Each point is the average difference between the station magnitudes of DJR and UB0 for all events listed in the ISC catalog, plotted as a function of the UB0 magnitude.
57 2.7 Discussion - The results described above provide convincing evidence that instru-mental clipping of analog stations is an important problem, and that the magnitudes of larger events published in the ISC catalog are biased lov and unreliable. A ccrollary to this conclusion is that it is virtually impossible to study the seismicity characteristics of different regionc using this (or sLsilar) catalogs, since each region is " monitored" by a different set of stations, with different operating and reporting character-istics. The VEIA arrays appear to be unique in their vide dynamic range, and, until a global network of digital stations becomes available and has accumulated a substantial data set, the VELA data is the only reliable source of information on upper bounds. So far, we have not discovered i any evidence for regicnal variations in seismicity using these arrays. As an example, Figure 20 shows data for shallow seismicity along the South American subduction zone. The global curve (Figure 9) is again an excellent fit. If we assume that the VELA seismicity curve is valid and represents saturation of the m scale, we can use similar arguments to those of b
. Chinnery and North (1975) to construct an ab -* **"t #'i"ti "8hi P' Assuming that the relationship between mb and M, at low magnitudes is m
b"M s+ 0.5 (2.3) (see, for exa=ple, Lambert et al, 19?h), then the form of the m -coment y curve is as shown in Figure 21. Some doubt about the constant in equation 2.3 remains, so the horizontal location of the m ~*O**"* "#V' b is not vell defined. n - --
,w-.e
4 i
- l 58 l
c22-57k8 1000 - _- SOUTH AMERICAN EVENTS 1966-70 ' 6 ob ; . UBO 0 6 O TFO
- O O A BMO 100 0 *
- . b O
o A o
]O : 0
~
.ee
_ . g O
^
O O - A
~
O O 1 I I I i i i.a.A 3.5 4.0 4.5 5.0 5.5 6.0 6.5 6.0 MAGNITUDE (mb ) $ Fig. 20: Frequency-magnitude data for South American events observed at 3 VELA ar:2ys. The aclid curve is the same as that in Figure 9, ad. justed vertically for a best fit.
p r j 59 E b
' j. i r
S I b R Y j5 n C22-5622 }. 5;- 9 M s ~ 8.6 o
?e N
\\ $
1 @ 8 - M3~ 7. 2 $
\ :
If
~
I 7 -
*b I tu mb ~ 7.3 r [
c ', 8 m b~ 5.8 f.1 f 56
<t
! I 2 Ms [L l 5j e
$ $g 5
- .j i:
- l. Y. 6 .
ir , i. 4 - 1 , '., i a I i & 1 I i i l l 1
', 7 23 24 25 26 27 28 29 30 31 22 q. ,
LOG (moment) o\
, p
'i 'N a '
,.' I.,
f! Fig. 21: An empirical =b-mement relationship consistent with the ;' F l VELA seismicity curve (Figure 9). "he M3 -moment relation- i ship from Chinnery and North (1975) is shown for comparison. l' H e s< ap i I h I i 1 1 l
.. . - rr -r ' ft
60 1 l The interpretation that the curve in,the VELA seisnicity relationship is due entirely +t saturation of the a scale seems reasonable. The , b shape of the m b
-m ment curve in Figure 21 is similar to that of the M,-
I moment relation, and (at least qualitatively) mbappears to saturate at about the expected magnitude. It therefore seems unlikely that any information about upper bound magnitudes can be obtained from the existing global m catalogs. ; b 2.8 Conclusions ,
~
The conclusions of this study are very negative. It does not ' appear that the best earthquake catalog data can shed any light on the problem of the existence or the regional variation of maximum earthquake size. This leaves only the much less comprehensive catalogs of Gutenberg and Richter (195h) and others, collected before 1960. While these older
~
catalogs are useful for event times and locations, there are growing j indications that the assigned magnitudes in these catalegs are unreliable (e.g. Chen and Molnar,1977). At least part of this unreliability probably arises from the instrumental pr,oblems described above. e s w e k
$ap %s
61 REFEREICES
# References marked by an asterisk are included for '
completeness, but were not used during this study. Many have not been translated into English. Aki, K. , Scaling law of earthquake source time-function, Geophys. J. , 21, 3-26, 1972.
. *Anan'in, I. V., Assessment of the seismic activity and the maximum possible energy of earthquakes in individual seisnogenic zones in the Caucasus region, in Seismogenic Structures and Seismic , Dislocations, VNII Geofizika, Moscow, 1973.
Archambeau, C., Estimation of non-hydrostatic stress in the earth by seisnic methods: lithospheric stress levels along the Pacific and Nazca plate subduction zones, manuscript.'in press, 1978. Bath, M., Earthquake energy and magnitude, in Physics and Chemistry of the Earth. Volume 7, Pergamon Press, 1966. Berg, J. W., Gaskell, R., and Rinehart, V., Earthquake energy release , and isostasy, Bull. Seism. Soc. Am., Sh, TTT-78h, 196h..
*Bonilla, M. G., and Buchanan, J. M., Interim report on vorldwide historic surface faults, Open file report, NCER, U. S. Geological Survey, '
1970. Borisov, B. A., and Reysner, G. I., Seismo-tectonic prognosis of the ; maximum magnitude of earthquakes in the Carpathian Region, Izvestia. Earth Physics, no. 5, 21-31, 1976.
*Borisov, B. A., Reysner, G. I. and Sholpo, V. N., On the preparation and use of geological-geophysical data for the identification of zones
, with different Mmax values in the outer zone of the Alpine folded region, Symp. on Search for Earthquake Predictors (Abstracts), MGGGS, MASFZN, Tashkent ,197h. Brazee, R. J. , Further reporting on the distribution of earthquakes with respect to magnitude mb, Earthquake Notes, h0, h9-51, 1969 i Brazee, R. J., and Stover, C. W., The distribution of earthquakes with respect to magnitude mb, Bull. Seism. Soc. Am.,}],,1015-1017,1969 Brune, J. N., Seismic moment, seismicity, and rate of slip along major i fault zones, J. Geochys. Res., 73, TTT-78h, 1968.
*Bune, V. I., Kirillova, I. V., Anan'in, I. V., Vvedenskaya, N. A.,
Reysner, G. I. , and Sholpo, V. N. , Attempt to estimate the maximum seismic risk: example of the Caucasus, Problems of Eng. Seism. no. Ik, Science (Nauka) Press, Moscow, 1971. ' W ' - - p-e-g-- 'w -.,ya vy-..w-m-r .,s w,ew wv, +
-w-rw. yw ev e-+y-v -w*w--ww y- y,,y,--rpy---o, --
w gy,wm y-w e---t w- r w re-em - --
l 62 I 3 - < -
*Bune, V. I. , at.d Polyakova, T. P. ,_ Correlation of the maximum possible earthquakes in the Caucasus region and Asia' Minor with seismic activity, in Investigation of Seismic Conditions, "Stiintsa",
Kishinev, 197h.
*Bune, V. I., Turbovich, I. T., Borisov, B. A., Gitis, V.'G., Reysner, G. I. , and Yurkov, E. F. , Method of development of a relationship - s between the earthquake magnitude and the tectonic parameters of '
a region, Proc. Acad. Sci. USSR, 21h, 197h. , Bune, V. I., Turbovich, I. T., Borisov, B. A., Gitis, V. G.,'Reysner,' G. I., and Yurkov, E. F., Method of prognosticating the maximum magnitude of earthquakes, I vestia, Earth Physics, no.10, 31 h3,1975 - Caputo, M., A mechanical model for the statistics of earthquakes, magnitude, moment and fault distribution, Bull. Seism. Soc. Am., 6],, 8h9-861, 1977 Chen, W.-P., and Molnar, P., Seismic. moments of major earthquakes and the' average rate of slip in Central Asia, J. Geophys. Res., 82,, 29h5-2970, 1977 4 Chinnery, M. A., Theoretical fault models, Pubs.' Dom. Obs. Ottawa, 2_7,, 211-223, 1967 Chinnery, M. A., and Lacoss, R. T., Magnitude differences between-station pairs, in Seismic Discrimination, Semi-Annual Technical Sinnury, Lincoln Laboratory, M.I.T. , 30 June 1976. { Chinnery, M. A., and North, R. G., The frequency of very large earth-quakes, Science, 190, 1197-1195. 1975 Chinnery, M. A., and Rodgers, D. A., Earthquake statistics in Southern New England, Earthquake Notes, h , 89-103, 1973 Christoffersson, L. A., Lacoss, R. T., and Chinnery, M. A., Statistical models for magnitude determination, in Seismic Discrimination, Semi-Annual Technical Summary, Lincoln Laboratory, M.I.T., 31 December 1975 . , , Connell, C. A., Engineering seismic risk analysis, Bull. Seism. Soc. Am., l 5_8_, 8 1583-1606, 1968. ,
,Connell, C. A., and Merz, H. A., Seismic risk analysis of Boston, ASCE National Structural Engineering Meeting, Cincinnati, Ohio, a
pril 1974 Connell, C. A. , and Merz, H. A. , Seismic risk analysis of ' Boston, l J. Struct. Div., ASCE, 101, no. ST10, 2027-20h3, 1975 Cosentino, P. , Ficarra, V. , and Lucio, D. , Truncated exponential , frequency-magnitude relationship in earthquake statistics, Bull. !_ Seism. Soc. Am., 61, 1615-1623, 1977 s . . . , ~ h c ,,
)
. -. ... .- . . - - . - - ~ -. . . -
.63 4
Cosentino, P., and Luzio, D.,'A gene,ralisation'of the frequency-magnitude relation in the hypothesis of a maximum regional magnitude, Ann. Geofis. (Rome),22,1-2,1976. Davies, G. F., and Brune, J. N., Regicnal and global fault slip rates from seismicity, Nature,' 229, 101-107, 1971. Drumya, A. V., and Stepanenko, N. Y., Map of the maximum possible. earth- . quakes of the 'Vrancea seismic' region, I:vestia, Earth ~ Physics , ' no.10,
.. 77-78,'1972.
Duda, S. J., Secular seismic energy release in the~ Circum-Pacific belt,
. Tectonophysics, g, h09 h52, 1965 "D:hibladze, E. A., Seismic activity and the maximum earthquakes in the Territory of Georgia and its vicinity, in Study of Seismic Danger, " Fan", Tashkent, 1971.
Epstein, B. , and Lommitz, C. , A model for the occurrence of large earth-quakes, Nature, 211, 95h-956, 1966.
*Esteva, L., Seismicity prediction: a Bayesian approach, Proc. hth W.C. E. E. , 1,, Santiago , Chile, 1969 Esteva, L., Seismic risk and seismic design decisions,.in Seismic Design for Nuclear Power Plants, M.I.T. Press, 1970.
l Evernden, J. F. , Study of regional seismicity and related problems, Bull. Seism. Soc. Am., 60, 393 kh6, 1970. Gayskiy, V. N. , and Katok, A. P. , The application of the theory. of extreme values to the problems of. recurrence of large earthquakes (in Russian), in Dynamics of the Earth's Crust, AN SSSR, Nauka, Moskov, 1965. Gelfand, I. M., Guberman, S. A., and- eilis-Borok, V. I., Pattern recognition applied to earthquake epicenters in California, J. Phys. Earth Plan. Int., 11, 227-233, 1976. Gorbunova, I. V. , A maximun-earthquake map for the Northern Tien Shan, I:vestia, Earth Physics, no. 11, 3-lh, 1969 Gambel, E. J., Statistics of Extremes, G:1umbia Univeraity Press, 1958.
. +
Gutenberg, B. , and Richter, C. F. , Seismicity of the Ear .h ar.d ?alsted Phencmena, Princeton University Press, 195h. 1 Eadley., J. F., and Devine, J. F., Seismotectonic map of the' Eastern 2 U. S. , U. S. Geological Survey, Misc. Field Study BT-620,197h.
" Hamilton, D., Final safety analysis report, Diablo Canyon Ruclear Power Plant, Appendix 2.5D, A=endment 19, p. 2.5D63-65, 1975. ;
-l l
~
1 I i 'l - m ,n
< v --
w -
6h
" Hor = ann, R. 3., State of the art for assessing earthquake hazards in the (Jnited States, Rept.- No. 3, U. S.. A. / Eng. Waterway Exp. Station, Vicksburg, Miss., Misc. Paper S-73-1, 197h.-
Housner, G. W. , Design spectru=, in Earthquake Engineering, ed. R. L. Wiegel,. Prentice-Hall, Inc., 1970.
*Kallaur, T. I., Seismic activity and the energy of =axi=um earthquakes in some regions of Tark=cnia, in Study of Seis=le Danger, " Fan",
Tashkent, 1971. Kanamori, H., and Anderson, D. L.', Theoretical basis for some empirical , relations in seismology, Bull. Seism. Soc. Am., g , 1073-1095, 1975 Kana=ori, H. , and ' Cipar, J. J. , Focal process of the great Chilean Earthquake, . May 22, 1960, Phys. Earth Plan. Int., 1, 128-136, 197k. Kana=ori, H., and Jennings, P. C., Determination of local magnitude ML from strong-motion accelerograms, Bull.~ Seism. See,. Am., @ , 471 h85, 1978.
- Karnich, V., and.Hubnerova, Z., The probability of occurrence of largest earthquakes in the European area, Pure and Appl. Geophys., 70,, ,
61-73, 1968. Knopoff, L. , and Kagan, Y. , Analysis of the theory of extremes as applied-to carthquake problems. , J. Geophys. Res. , 82,, 56LT-5657,1977 Kogan, L. A., and Shakirova, T. D., Assess =ent of K=ax by means of Gumbel's first =axi=um distribution, in Problems in the Assessment of the Seismic Danger, "Nauka", Moscov, 1974 Lambert, D. G., Tolstoy, A. I., and Becker, E. S., Vela Netverk Evaluation and Autematic Processing Research, Technical Report No. 7, Texas Instruments Inc., 9 Dece=ber 197h. I 1 ( Lee, V. N. K., and Brillinger, D. R., A preliminary analysis of the Chinese ! earthquake history, paper presented at the U.S. Geological Survey i Conference on Seismic Gaps, Boston, May 1978. Lo:nitz, C., Statistical prediction of earthquakes, Rev. Geophys., h, 377-394, 1966. McGarr, A., Upper limit to earthquake size, riature, 262, 378-379, 1976. McGuire, R.-K., Methodology for incorporating pars =eter uncertainties into seismic hazard analysis for low risk design intensities, - presented at Int. Sy=p. on Earthq. Struct. Eng., St. Louis, August , 1976. , Merz, H. A., and Connell, C. A., Seismic risk analysis based on a quadratic i magnitude-frequency law, Bull. Seism. Soc. A ,, 6J, 1999-2006, 1973 l i l I I
----.--._---------_--,_,.-.---_u_--.--__--_----__--__--___-----,,.--s--,-
65
'Milne, W. G. , and Davenport, A. G. , Earthquake probability, Pubs. Sec .
Obs. Ottava, Seism. Series, 1968-L, 19pp., 1968. Neunhofer, H. , non-linear energy frequency curves in statictics of earthquakes, Pageoph, 22,,76-83,1969 Nev= ark, N. M., and Rosenblueth, E., Fundamentals c' n"*" quake Engineering, Prentice-Hall Inc. ,1971. Nordquist, J. M. , Theory of largest values applied to earthquake
=agnitudes, Trans. Am. Geophys. Union, 96, 29-31, 19L5.
North, R. G. , Station Magnitude Bias , Its teter=ination, causes and F"fectc , Lincoln Laboratory, M.I.T., Technical Ucte 1977-2b, 1977 Otsuka, M., Cut-off of seismic energy, J. Phys. Earth, 21, 119-123, 1973 Papazachos, B. C., Dependence of the seis=ic para =eter b en the tagnitude ! range, Pageoph, 112, 1059-1065, 197L. P'ei-shan,- C. , and Pang-Lui, L. , An application of statistical theory of extreme values to moderate and long-interval earthquake prediction, Acta Geophys. Sinica, lj[, 6-2L,1973 (Plenum Publishing Corp. translation,1975).
*Perkins , D. , The search for taximum =agnitude, 30AA Eartha. Inf. Bull . ,
18-23, July 1972.
*Putrushevskiy, 3. A. , On the relationship between the earthquake of maxi =um intensity and the geological state, Bull. Council Seism. ,
no. 8, 1960. Richter, C. F. , Elementary Seismology, W. E. Freeman and C;:pany,1953. Rikitake, T., Statistics of ultimate strain of the earth's crust and probability of earthquake occurrence, Teetenochysics, 20,, 1-21, 1975.
. Ringdal, F. , Maxi =um-likelihood esti=ation of seismic =agnitude, Bull.
Seism. Soc. A=., 6_6,, 789-302, 1976.
, Ri:nichenko, Y. V. , P ssibilities of calculating maximum earthquakes, Trans. (Trudy) Inst. Phys. Earth. Acad. Sci. USSR , 21, 192, 190 2.
Ri:nichenko, Y. V., Relationship between the energy of tha str:ng at earthquakes and seismic activity, 20k111:' 0:ad. Usuc 53SR, 157, 1352-135L, 196La. Ri:nichenko, Y. V. , Determination of the energy flux of earthquake foci on the basis of seismic activity, Ocklady Akad. Neuk S3SR, 159, 321-322, 196Lb.
, g g 4 g 9 'A* 3
*' .. 7NO MD '*# # . P
66 "Ri:nichenko, 1. V., Seismic activity and'the energy of the largest earthquakes, in Problems of the Geochysics of Soviet Central Asia KassAtstan, Science (Nauka) Press, Moscow, 1967.
"Ri:nichenko , Y. V. , The generalized law of earthquake occurrence, Boll. di Geofis. Theor. ed. A;olic., 12, no. L8, 1970. *Ri:nichenko, Y. V., The strongest possible earthquakes, Earth and Universe (Zemlya i Veslennaya), no. 5, 1971. "Ri:nichenko, Y. V., Determination of seismic danger, in Problems in the Quantitative Assessment of Seismie Danger, "Nauka", Moscow,19Th.
Ri:nichenko, Y. V., and Bagdosarova, A. M., The strongest possible earth-quakes of Japan, I:vestia, Earth Physics , no.11,1h-32,1975 "Riznichenko, Y. Z., Dru=ya, A. V., Stepanenko, N. Y., Crustal seismic activity and the maxi =u= possible earthquakes in the Carpathian-Balkan Region, in Regional Studies en Seis=ie Reci=e, Shtinitsa, Kis hinev, 1974. Ritnichenko, Y. V. , and Dzhiblad:e, E. A. , Determination of the rav4 mu= possible earthquakes on the basis of the ec=prehansion data of the Caucasus Region, I:vestia, Earth Physics, no. 5, 6L-85, 197h. Ri:nichenko , Y. V. , and Zakharova, A. I. , Generalized law of earthquake occurrence, I:vestia, Earth Physics, no. 3, 29-38, 1971. Rothe, J. P. , The Seismicity of the Earth 1953-1965, UNESCO Publication, 1969 Shakal, A. F., and Toksoz, M. N., Earthquake hazard in New England, Science, 105, 171-173, 1977 Shebalin, N. V., The =aximu= =ag.itude and maxi =':= scale intensity of an earthquake, I:vestia, Earth Physics, no. 6, 12-20, 1970. Shebalin, N. V. , Esti=ation of the size and position of the focus of the Tashkent earthquake frc= =acroseismic and instrunental data, in The Tashkent Earthquake cf 1966, Uzbek Branch, Acad. Sci. USSR Press, 1971. S,hebalin, N. W., Assessment of the taxi =un seismic danger in the Crimea-Ta ansk region, in Seis=icity, Seis=ic C1nger in the Crimea and Seis=ostability of Structu es , "naukovaya Ot=2a", Kiev, 1972. Shenkova, Z. , ani Karnik, V. , The probability of occurrence of largest ear;hquakes in the European area - Part II, Pure and A ;1. Geophys. , 80, 152-161, 1970. I 1 Shenkova, 2., and Karnik, V., Cenparison of metheds of determining l the largest possible earthquakes, I vestia, Isrth ?hysics, no. 11, . 118-125, 197h. 1 i
67-Shlien, S.,'and Tokso:, M. N., Frequency-magnitude statistics of earth-quake occurrence, Earthquake Notes, bl, 5-18, 1970. Smith, S. W., Determination of maximum' earthquake magnitude, Geophys. Res. Letters, j;, 351-35h, 1976. Stewart, I. C. F., On the use of the maximum likelihood estimator j
- for recurrence curves, Earthquake Notes _, hi, 17-22, 197h.
*Tsuboi, C., Isostasy and maximun earthquake energy, Proc. Imp. Acad.
Japan, 16,, 19k0. l i
. Tsuboi, C. , Earthquake energy, earthquake volume, aftershock area and strength of the eerth's crust, J. Phys. Earth, h, 63-66, 1956.
i
*Uspenskaya, T. A. , Experience obtained in calculating the map of maximum possible earthquakes in the Fribaikal region, in Study of Seismie Danger, " Fan", Tashkent, 1971. !
Veneziano, D., Probabilistic and Statistical Models for Seismic Risk Analysis, M.I.T. Dept. of Civil Eng., Publication R75-34,-1975 Von Seggern, D., Joint magnitude determination and analysis of variance-for explosion magnitude estimates, Bull. Seism. Soc. Am. , 63., 827-8h5, 1973 Yegulalp, T. M., and Kuo, J. T., Statistical prediction of the occurrence of maximum magnitude earthquakes,. Bull. Seism. So" Am., 6h_, 393-bih, 197h.
*Zakharova, A. I., Computer program for calculating Kmax maps, in Investigation of Seismic Conditions, "Stiintsa", Kishinev, 1974.
*Zonin, Y. A., and Novoseleva, M. P... Prediction of long-term seismic activity on the basis of the geomorphological and geophysical parameters of the Pribaikal region, in Problems in the Quantitative Assessment of Seismic Danger, "Nauka", Moscow, 1974.
l. 1 p& te 4g. g .asy m.gq pg e a semes ,.m f re p4 e p pfy.W e64- *4 M W "D.JS 98 'M MM ' { 3 , ,M
- 4 %8499W4 ?y 4 % 9P ONW ,-_ ,yu y ,. y .
60-r APPENDIX- .
~
Progress Reoort: New England Crust and Upper Mantle' Structure The recent establishment of the northeastern seismic array has allowed us'to construct a preliminary model of the crust and upper mantle structure beneath New England. Because the array has only been i in full operation for approximately 2 years, the dataset is itmited, and we have analyzed the data using a vaciety of techniques including:
- 1. observations of relative JB residuals (
- 2. a time term analysis using P arrivals 3 three-dimensional modeling using teleseismic P-vaves
- h. analysi.e of array diagrams 5 refraction studies ,
Preliminary results indicate a crustal thickening under central New Hampshire coupled with a slight crustal thickening vestward towards the North American craton. There is also some suggestion of a region of relatively low velocity in the upper mantle beneath central New Hampshire and southern Maine. Methods of Analysis and Results The relative arrival times of teleseismic P vavs$ vere read from l enlarged copies of 16 mm develocorder film. In general, the first few cycles exhibit coherence across the array so relative arrival measurements were taken from a prominent peak or trough early in the signal. This procedure was required for a number of weakly recorded teleseisms in which the first break was too emergent or obscured by noise. In this vay, arrival times could be measured to 0.1 sec. Elevation corrections ' vere applied to the data by assuming a vertical phase velocity of 6.0 kn/sec and dividing this into the station elevations. ! b ,
. . = . . _ _ . - -~
?
69 Absolute travel time residuals were calculated with respect to JB
- r tables and are defined to be JB _ , obs , JB E
ij ~ 'ij ~ 'ij vhere R g is the absolute residual with respect to JB tables for bs station i, event j ; T is the observed travel time using origin > gJ times from PDE bulletins; T g is the theoretical travel time through a JB earth. The residuals were reduced by calculating relative residuals with respect to a mean residual computed for each event; > R =R 1 [ R ij ij ij N , vhere H is the number of stations reporting P arrivals for a given event. The utilization of relative residuals reduces source effects and mislocation errors, removes errors in origin time, and reduces effects of travel path through an inhomogeneous mantle. In this way, positive residuals represent late arrivals where the waves have been slowed in the crust or upper mantle beneath the array. There are several consistent trends in the teleseismic P vave residuals which suggest the presence of large scale regional structures ? in the crust and upper mantle beneath the array. The data show both 9
. azimuthal variations in residual values, and variations in average station residuals across the array. ;
The data were inverted to a depth of 350 km using the three dimen-sional medeling technique of Aki et al. (1977). Perhaps the most inter-esting result is the presence of a regicnal zone of relatively low velocity in the upper mantle beneath central new Hampshire and southern , Maine. This zone of relatively low velocity correlates spatially with the Meso. 4 e White Mountain plutonic series. It is thought that the e
i 70 i scurce of these intrusive complexes is deep-seated (Chapman, 1976), and it is possible that this anomaly is related to the formation of these ( plutons. ! i S time term analysis using P arrivals indicates that the variations ! in average station residuals may be due to variations in crustal thickness and/or velocity. This is in contrast to the observed azimuthal distribu-tion of residuals for each station which is probably due to deeper effects. It was assumed that the distribution of average residuals is caused by crustal thickness variations, and the data were inverted to F find a crustal thickness map of New England. The resulting map su6 gests a crustal thickening beneath central New Hampshire, with more normal thicknesses in Massachusetts and Maine. The contours of the map parallel the northeasterly trend of the Appalachians. The variations in crustal thickness observed across the network are also supported by analysis of array diagrams. These are stereographic projections of slowness and asinuth anomalies observed from a plane wave fit to the wavefront traversing the network. These studies indicate a l Moho which dips 2* or less to the northwest. This is not surprising because it is expected that the crust would thicken from the continental margin towards the North American craton. In addition to the above mentioned studies, an average crustal l velocity model has been compiled for eastern Massachusetts and southern New Ha=pshire by combining results from timed quarry blasts with the time term analysis. The model is currently being used in earthquake location progra=s at M.I.T. and is as follows: l i i .
l l 71 ) I layer (km) h elocity (km/sec) 0 - 7.3 5.68 7.3-26.1 6.26 26.1-38.0 7 33
, Moho 8.13 Future Studies Studies for the next year vill be aimed at improving the preliminary crust and upper mantle model for New England. This vill be achieved by
! using additional teleseismic P and PkP data. The database is currently being expanded to include readings from short period stations in Connecticut i and eastern New York. I i The structureel models derived from the residual studies vill be l compared to those from long period surface wave dispersion studies, i l Phase velocities are presently being computed as a function of azimuth from the Qaebec-Maine border event of June 15, 1973, and simple crustal models will be developed. Phase velocities vill also be measured using the two station technique. More elaborate models vill be generated by performing a simultaneous inversion of phase velocity and attenuation following the techniques of Lee and Solomon (1975). l A study of the Lg phase, a short period higher mode Love wave, vill be initiated to compare the effect of regional geologic structure on Lg propagation. The data vill be collected using three component, digital recording event detectors developed at MIT. tma
- . - - - . __ . - ~. - .
72 References , , i Aki, K., A. Christoffersson, and E. S. Husebye, Determinatiot: of the three-dimensional seismic structure of the lithosphere, L. Geophys. Res, 8,2_, 277-296, 1977 , Chapman, C. A. , Structural evolution of the White Mountain magma series, Geol. Soc. _Am_, Mem., lh6, 281-300, 1976. Lee, W. B. sud S. C. Solomon, Inversion schemes for surface wave attenua-tion and Q in the crust and the mantle, Geophys. 1. R_. Astron. Soc., h_1,h7-71,1975 i e 4 i l D e
-- l
- . -~ __.
4 E::hibit 4 1 i 4 1 ) 1 1
- 1 1 -e-w x- . - ,, e + r , - , , - - - -r -- n w = -,ee ,- - ,,m-w .., .,-m-- a e v , -,- < - - -- - - . + , --en---+ , , - - - r,
RESUME PERSONAL DATA , Name: Michael A. Chinnery [ Date of Birth: 27 September, 1933 t Place of Birth: London, England Citizenship: U.S. - ;
- Marital Status: Married, two daughters Wife's Name: Thora Elizabeth (nee Hawkey) -
Wife's Place of Birth: Peterborough, Ontario, Canada i Wife's Nationality: Canadian ; Military Service: Royal Air Force, 1952-5h Pilot Officer (Flying Officer, Beserve)
~
Rank: Branch: Fighter Control Area: England and Egypt Present Home Address: 110 Gray Street Arlin6 ton, Massachusetts ' 0217h U.S.A. Present Business Address: Lincoln Laboratory, M.I.T. h2 Carleton Street Cambridge, Massachusetts 021h2 U.S.A. Telephene: Home 617/6h6-0937 Business 617/253-7352 Security Clearance: Secret
, 'i
/ / ,
,[
DisTE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . S IG I!ED . . . J. . . . . . '.' . . . . . . . . . .....,. , o I-h n I ,l' *
.-,4,,.. , , . . . . ,. . . . .. . . . v.,y. . . , , , , ,..,--,...,..,..,..y. , . , .,~c., . . ,, . ,,.,_,g-,w..,_ .,s, ..,,,.4.,.,,,,,,._,,,,%,,,y , , , , , , ..,._,m,7.;, ,,.%,.,,% ,,og c..y.. ..w.,,.-,,,,.. 9 -,g
. )
i i EDUCATION . High School: Brentwood School, Brentwood, Essex, England 1944-51 Praepostor 1949 School Praepostor 1951 Head of House 1951 County Major Scholarship 1950 State Scholarship 1951 - Undergraduate: Corpus Christi College, Cambridge 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 (Physics) 1957 (upper 2nd Class) Graduate: Geophysics Laboratory, Department of Physics, University of Toronto, Canada 1958-62 ' Computation Center Fellowship 1959 Canadian Kodak Fellowship 1960 Imperial Oil Fellowship- 1960 . National Research Council Fellowship 1960 M. A. Thesis: "The App 1tcation of Dislocation Theory to Geodynamics" (Advisor: J. A. Steketee) Ph.D. Thesis: "The Dynamics of the Strike-Stip Fault" (Advisor: F. S. Grant) Degrees: A.V.C.M. 1945 Victoria College of Music B. A. 1957 bambridge University M.A. 1959 University of Toronto M.A. 1961 Cambridge University Ph . D. 1962 University of Toronto M. A. (ad eundem) 1967 Brown University Sc.D. 1977 Cambridge University L - . . . _ _ _ . . . . _ _ -
i EMPLOYMENT 1 Trainee Engineer (computer construction and development) Plessey Company, Ilford, Essex, England 1954 (summer) l 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 1958 (summer) Geophysicist (fleid party 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 . Dept. of Earth and Planetary Sciences, M.I.T. 1965-66 Associate Professor Dept. of Geological Sciences Brown University Providence, Rhode Island 1966-71 Professor Dept. of Geological Sciences, Broivn University 1971-73 Senior Research Associate Dept. of Earth and Planetary Sciences, M.I.T. 1973-present Group Leader Applied Seismc, logy Group, Lincoln Laboratory, M. I.T. Cambridge, Massachusetts 1973-present m -- - - , - w , n -
-m
.... . . . .. _ - .._ .... - - - ~ . - . . .. ._-... . - .
l i a . . e CONSULTANT TO 2
.i Huntec Lt'd, Toronto -
1958-65' l Arthur D. Little, Inc, Cambridge, Massachusetts - 1966-73 i Earth Sciences Research, Inc, Cambridge, Mass. 1969-73 i Lincoln Laboratory, M. I.T. 1971-73 National Aeronautics and Space Administration ~ 1976-present > (Lincoln Laboratory does not allow its employees to consult for Industry) FIELD WORK r Seismic exploration,- Milford Haven, England 1957 ! Electromagnetic, resistivity, magnede and se!smic ! exploration in Alaska, Northwest Terrttories, l and Alberta 1959 ' Gravity survey, British Columbia 1960 Shallow seismic exploration, Northcrn Quebec 1961 ) 1 Gravity survey, Northern Ontario ; 1962 Shallow selsmic exploration, British Columbia 1964 COURSES TAUGHT i Applied Geophysics (physics undergraduates) i Applied Geophysics (geology undergraduates) l Elasticity Theory (graduate) Dislocation Theory (graduate) Introduction to Geophysics (undergraduate / graduate) L Introduction to Seismology (graduate) Earthquakes (introductory undergraduate) f Planetary Physics (undergraduate) Data Analysis (graduate) Tectonophysics (graduate) plus various seminars and portions of courses + w se --r,-e+- e,- e nei a 4 +,-c.., em-g-.,,+,.,., ,p ,se, ,wm , +.py gewm.,,+,, y a y -g gys ,e - # y ,m gi ,-g++ge:.wi ,-ae&-,,,y.-+ p sev ,,, .g. , w- e -w ee es g+g- e 9- +sep9,,--,w-3r., y,y--eyy c+ ew -s yw
3 - 3
$ b PROFESSIONAL SOCIETIES AND OFFICES HELD American Geophysical Union-Secretary, Tectonophysics Section, 1968-70 .
Program Chairman, Tectonophysics Section,1969 Annual Meeting Program Chairman, Tectonophysics Sectioni 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 j Secretary, Vancouver Center,1963 i 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 1
Royal Astronomical Society Fellow,1973-present 4
- presently inactive l
I
_. _ ._ . . __. _ _ - _ . _ _ _ _ ~ . . _ COMMITTEES AND MISCELLANEOUS ACTIVITIES . Resident Faculty Advisor, Acadia Residence, University of British Columbia, 1962-64 Member, Gravity Sub-committee, National Research Council of Canada, 1964-65 Associate Resident Fellow, Mead House, Brown University, 1967-69 4 Member, Dining Services Committee, Brown Univers[ty, 1969-71 Member, Graduate Council, Brown University, 1969-71 Chairman, University Lectureships 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) Lecture series (chairman) Geophysics committee (chairman) ,
^
Undergraduate program committee (member) < Graduate admissions and awards commi' tee (chsdrman) Testified before the Advisory Committee on Reactor Safeguards, Nuclear Regulatory Commission, concerning seismic risk at the Seabrook nuclear power plant , site,1974 .- Appeared as expert witness at the licensing hearings for the Seabrook nuclear power plant,1975 Member, Panel on Seismograph Networks, Committee on Seismology, National Academy of Sciences, 1975-77 . 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 Meeting Chairman, Summer workshop on the application of space techniques to geodynamics, N. A.S. A. , Denver,1977 Member, Working Group on Upgrading WWSSN Stations, National Academy of Sciences,1977 Gave special invited lecture on the appilcation of space techniques to geodynamics, International Association of Seismology and Physics of the Earth's interior, Durham, England,1977 Member, Panel on Storage of Digital Seismic Data, Committee on Scismology, National Academy of Sciences, 1977-78 h -v c .. -- - c --vw +g.,,-- --- ,- . . . - , .wa+ .--,--c.-y, er = ,,-,--&-,,-re-+.<s.ve% .,--v e w v w. e % .=r ,-.e t *w- -,+--g-4-+.<,raew-os-r e--:e m r n r-r---+---,-r'
Member, Working Group on Solid Earth Data,. Committee on Data Interchange and Data Centers, National Academy of Sciences, 1977-78 Member, Proposal review panel, N. A.S. A. ,1978-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
- 1. A. S. P. E. I. representative to joint I. U. G. G. /1. 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, Aestralia,1979 Gave technical presentation to the Nuclear Regulatory Commission on the application of probabalistic methods to the estimation of seismic risk, Washington,1980 Member, Panel on Data Problems in Seismology, Committee on Seismology, National Academy of Sciences, Woods Hole,1980 !
l 1 l l l l l
~ . _ _ . _. . _ . . - _ . . _ __
PUBLICATIONS The following list includes a* variety of.different kinds of publications. Papers in scientific journals are indicated by an asterisk (*).
- 1. Chinnery, M.A., The application of dislocation theory to geodynamics, M.A. thesis, University of Toronto, 88pp., 1959
- 2. "Chinnery, M. A. , Some physical aspects of earthquake mechanism, J. Geophys. Res., 65, 3852-54, 1960. '
3 "Chinnery, M.A.,' Terrain corrections for airborne gravity gradient measurements, Geophysic[, 26,, h80-89, 1961.
- h. *Chinnery, M.A., The deformation of the ground around surface faults, Bull. Seism. Soc. Am., 51, 355-72, 1961.
5 Chinnery, M. A. , The dynamics of the strike-slip fault , Ph.D. thesis, University of Toronto, 138pp., 1962.
- 6. "Chinnery, M.A., The stress changes that accompany strike-slip faulting, Bull. Seism. Soc. Am., 23.,921-32,1963.
7 *Chinnery, M. A. , The strength of the earth's crust under horizontal shear stress, J. Geophys. Res., 69, 2085-89, 196h.
- 8. "Chinnery, M.A., The vertical displacements associated with trans-current faulting, J. Geophys.# Res. , 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, h18-2h, 1965
- 10. *Chinnery, M.A., Secondary faulting I: Theoretical aspects, Can. J. Earth Sci., 3, 163-7h, 1966.
- 11. *Chinnery, M.A., Secondary faulting II: Geological aspects, Can. J. Earth Sci . , 3, 175-90, 1966.
- 12. Toksoz, M.N., and Chinnery, M.A., Seismic travel times from Long-shot and structure of the mantle (abstract), Trans. Am.
Geophys. Union, k7, 16h, 1966. 13 Chinnery, M. A. , The dislocation fault model with a variable discon-tinuity (abstract), Trans. Am. Geophys. Union, h7, 166, 1966, 1h. "Chinnery, M.A., and Toksoz, M.U., P-wave velocities in the mantle below 700 km, Bull. Seism. Soc. Am., 57, 199-226, 1967 15 *Chinnery, M. A. , Leor.bruno , W. , McC:nnell, R.K. , and O' Brien , J. , ' Observations of fatigue deformation in contact loading, ASFE publication 67-DE-53,12pp. ,1967 ;
'e j- 5
.. _ ~ -. . - - . . -
i
- 16. Chinnery, M. S. , The vertical displacenents associated with ,
transcurrent faulting, in Proceedines of the VESIAC-conference on the current status and future trognosis for understanding the source mechanise of shallow seismic '{ events in the 3 to 5 narnitude range, VISIAC report 7885-1-X, 299-306, 1967
- 17. *Toksoz, M. N., Chinnery, M. A., and Anderson, D. L., Inhomo-geneities in the earth's mantle, Geonhys. J., 1_3_, 31-59, 1967
- 18. Chinnery, M. A., Evidence for lateral variations in the lover mantle (abstract), Trans. Am. Geothys. Union, h,8_, 19h, 1967 19 Chinnery, M. A., Source time function for a vrench fault move-ment (abstract), Trans. Am. Geophys. Union, h_8, 8 203, 1967
- 20. Chinnery, M. A. , Theoretical investigations of the mechanism of faultir.g, in U.S. Uoner Mantle Proiect Progress Report, 126-7, 1967
- 21. *Chinnery, M. A., and Petrak, J. A., The dislocation fault model with a variable discontinuity, Tectonophysics, j,, 513-29, 1968.
- 22. Chinnery, M. A., and Rodgers, D. A., The stressed zone at the lover edge of a strike-slip fault (abstract), Trans. Am. Geoehys.
Union, M , 299, 1968. 23 Chinnery,M.A.,Earthquakemagnikudeandsourceparameters (abstract), Earthouake Notes, 3,?.,13,1968. 2h. Chinnery, M. A., Measurement of the first and second derivatives of the travel time curve using LASA (abstract), Geol. Soc. Am. Special Paner 101, 29h, 1968, 25 Chinnery, M. A. , Direct measurement of the second derivative of the travel ti=e curve (abstract), Geol. Soc. Am. Scecial Paper 115, 215, 1968. .
- 26. *Chinnery, M. A., Velocity anomalies in the lover mantle, Phys.
Earth & Plan. Int., 2_, 1-10, 1969 27 *Chinnery, M. A. Theoretical fault codels, Pubs. Dem. Obs. Ottava. 37., 211-23, 1969
- 28. Chinnery, M. A., Review of "Non-elastic Processes in the Mantle", I edited by D. C. Tozer, Trans. Am. Geothys. Union, 50,, h97, 4 1969 29 Rodgers, D. A. , and Chinnery, M. A. , The displacements and strains associated with a curved strike-skip fault (abstract),Trans.
An. Geothys. Union, JC_, 233, 1969
C .#
- 30. Chinnery, M. A. , The velocity an caly at 2000 km depth (abstract),
Trans. Am. Geochys. Union, 52, 2hh,1969
- 31. *Chinnery, M. A. , Earthquake magnitude and source parameters, Bull. Seism. Soc. Am., 59, 1969-82, 1969
- 32. "Chinnery, M. A., Earthquakes and the Chandler vobble, Cc==ents on Earth Sci. Geochys., 1, 1-7, 1970.
33 *Chinnery, M. A. , Earthquake displace ent fields, in Earthquake Displac ement Fields and the Rote. tion of the Earth, edited by Mansinha and others, Reidel Press, The Netherlands , 17-38, 1970. 3h. "Chinnery, M. A., The Chandler vobble, in Understanding the Earth, edited by Gass and others, Artemis Press, Great Britain, 89-95, 1971. 35 Rodgers, D. A., Chinnery, M. A., and McConnell, R. K., An assess-ment of the 6 1acial rebound techanism for earthquakes in the Eastern U. S. (abstract), Trans. Am. Geophys. Union, 12,277,1971.
- 36. Chinnery, M. A., and Jovanovich, D. B., The effect of the lithosphere-asthenosphere beundary on earthquake displace-ment fields (abstract), Trans. Am. Geochys. Uninn, }2., 3h3, 1971.
- 3T. Chinnery, M. A., P-vave arrivals at the Large Aperture Seistic Array, in Urner Mantle Projec+. U.S. Final Reoort, 56, 1971.
- 38. Chinnery, M. A. , Theoretical and field investigations of the mechanics of faulting, in Up;er Mantle Project U.S. Final
~
R ecort , 16k , 1971. 39 Chinnery, M. A., Investigations of F-vave arrivals at LASA, Final [ Technical Report, ARPA Contract Thh620-68-C-0082, 85pp,1971. ! Wells, F. J. , and Chinnery, M. A. , Variations in the annual component l h0. of polar motion at individual observatories (abstract), l Trans. Am. Geochvs. Union, 51, 3h5,1972. l 1 t'. A., Evidence for a cusp in the l bl. Jovanovich, D. B. , and Chinnery, travel time curve at 35 (abstract), Trans. Am. Geonhys. l Union, 13, h52, 1972. l I h2. "Chinnery, M. A., and Wells, F. J., On the correlation between earthquake occurrence and disturbances in the path of the rotation pole, in Rotation of the Earth, edited by Melchior and Yumi, Beidel Publishing Co., The Metherlands, 215-220, 1972.
h3. Rodgers, D. A., and Chinnery, M. A., A revised velocity structure for New En51 and (abstract), Trans. An. Geophys. Union, 53, h52, 1972. hh. " Rice, J. R. , and Chinnery, M. A. , On the calculation of changes in the earth's inertial tensor due to faulting, Geophys. J. , 22, 29-90, 1972. h5 *Chinnery, M. A. , and Jovanovich, D. B. , Effect of earth layering on earthquake displacenent fields, 3ull. Seism. Soc. Am., 62, 1629-1639, 1972. h6. "Rodgers, D. A. and Chinnery, M. A., Stress accumulation in the Transverse Ran6es, Southern California, In, Proc. Conf. on Tectonic Problems of the San Andreas Fault System, ed. R. L. Kovach and A. Nur, Geol Sci. ,13, School of Earth Sciences, Stanford University, 70-79, 1973 h7 " Wells, F. J. and Chinnery, M. A. , On the separation of the spectral components of polar motion, Geophys. J. , 3h,,179-192,1973 h8. Chinnery, M. A., Earthquake risk in Southern New England (abstract) Earthquake Notes, h , 29, 1973 h9 Chinnery, M. A., and Rice, J. R., On the calculation of changes in the earth's inertia tensor due to faulting (abstract), Geophys. J., 31, 373, 1973 "Chinnery, M. A. and Rodgers, D. ., Earthquake statistics in Southern 50. New England, Eartheuake Notes, hh,, 89-103, 1973
- 51. Chinnery, M. A. (Editor), Seismic Discrimination, 20th Semi-Annual Technical Su= mary, Lincoln Laboratory, M.I.T., 31 December 1973. .
- 52. Jovanovich, D. B., Husseini, M. I., and Chinnery, M. A., Displace-ment strains and tilt fields due to a point dislocation in a layered half-space (abstract), Earthquake Notes, h5, 15, 197h.
53 Chinnery, M. A., The International Seismic Month: Introduction (abstract), Earthquake Notes, h5, 2h, 197h. - 5h. Chinnery, M. A. (Editor), Seismic Discrimination, 21st Semi-Annual Technical Su= nary, Lincoln Laboratory, M.I.T. , 30 June 19Th. 55 Chinnery, M. A., and north, R. G., Frequency-magnitude curves and the M -r relationship, in Seismic Discrimination, Semi-Annual TSchicalSm-'ry,LincolnLaboratory,M.I.T.,30 June 197h, lh-15 r -- 7 n---sn
- - ._-_ - . .~ , - -_. .
i i E
- 56. Chinnery, M. A., and North, R. G.,.The Moment-M relationship, and '
the frequency of larse earthquakes, in Seismic Discrimination, Semi-Annual Technical Su= mary, Lincoln Laboratory, M.I.T.,
'30 June 197h, 15-16.
57 Chinnery, M. A., A Statement before the AIC Advisory Co=mittee on Reactor Safeguards, October 31, 197h, lipp.
- 58. *Jovanovich, D. B. , Husseini, M. I. , and Chinnery, M. A. , Elastic i dislocations in a layered half-space - I. Bacic theory and ,
numerical methods, Geophys. J. , }9,, 9 205-218,197h. 59 "Jovanovich, D. B. , Husseini, M. I. , and Chinnery, M. A. , Elastic disclocations in a layered half-space - II. The point source, Geophys. J., 3_9_, 9 219-2ho, 197h.
- 60. Chinnery, M. A. (Editor), Seismic Discrimination, 22nd Semi-Annual Technical Su==ary, Lincoln Laboratory, N.I.T. , 31 December :
197h.
- 61. Chinnery, M. A., Characteristics of short-ters variations in seismic activity, ,1,,n_ Seismic Discrimination, Semi-Annual Technical Summry, Lincoln Laboratory, M.I.T. , 31 December 1974, h9 !
r
- 62. Landers, T. E. , and Chinnery, M. A. , Spectral analysis of earth- '
quake occurrence rates, i_n_ Seismic Discrimination, Semi-Annual Technical Su m ry, Lincoln Laborator /, M.I.T., 31 December , 197h, h9-51. b a 63 Chinnery, M. A. and Landers, T. E., Short tern variations in the f' level of global seismic activity (abstract), Geological' Society of America Abstract with Pregrams,,7_, 3, 401, 1975 ( 6h. Chinnery, M. A..and Landers, T.'E., Evidence from earthquake time sequences for a larGc+ scale event involving the Pacific and ' Nazca plates during 196h-68 (abstract), T ans. Am. Geophys. Union,_56., 4h3, 1975 65 Chinnery, M. A. and North, R. G., Global frequency-magnitude relationships and their implications (abstract), Earthcuake Notes, h_6_, 55, 1975
- 66. Landers, T. E. , and Chinnery, M. A. , Spectral analysis of earth-quake occurrence rates (abstract), Earchcunke Notes, h_6,,,
55-6, 1975 6T. Chinnery, M. A., (Editor), Seismic Discrimination 23rd Semi-Annual Technical Su. mary, Lincoln Laboratory, M.I.T., 30 June 1975
- 68. Chinnery, M. A., Spatial and tenporal variations in the frequency-magnitude curve, _i,n_ Seismic Dicerimincion, Semi-Annual Technical '
Summary, Lincoln Laboratory, M.I.T., 30 June 1975, 55 e y e r.imy-. y w m.y-, g g...,. , ,,,,,,m.-e , ,.c ._-y,t-+ir+*m
- m- re---- gi v - 9 ww'-*:'--en vr w www e- 3 n --g fM*-1r-+--**-M-v *- e f*-'*
1
-69 Chinnery, M. A., and Landers, T. E., Correlations between seismic l netivity in videly separated regions, iIt Seismic Discrimina-tion, Semi-Annual Technical Si - "y, Lincoln Laboratory, M.I.*t.,
~
30 June 1975, 56.
- 70. "Cninnery, M. A. , The static deformation of an earth with a fluid core:
( a physical approach, Geochys. J., h2, h61 h76, 1975.
- 71. Chinnery, M. A. , and Landers, T. E. Short term variations in the level of global seismic activity (abstract), Earthcuake Notes, h6,, 26-7, 1975 ,
- 72. *Chinnery, M. A., and Landers, T. E., Evidence for earthquake triggering stress, Nature, 258, h90 h93, 1975
- 73. "Chinnery, N. A., and North, R. G., The frequency of very large earthquakes, Science, 190, 1197-8, 1975.
- 74. Chinnery, M. A., (Editor), Seismic Discrimination, 2hth Semi-Annual Technical S m ry, Lincoln Laboratory, M.I.T., 31 December 19T5 75 Chinnery, M. A., Problems in magnitude esti=ation, in. Seismic Discrimination, Semi-Annual Technical Su= ary, Lincoln Laboratory, M.I.T., 31 December 1975, 1-2.
- 76. Christoffersson, L. A., Lacoss, R. T., and Chinnery, M. A.,
' Statistical models for magnitude estination, in Seismic Discrimination, Semi-Annual Technical Sumry, Lincoln Laboratory, M.I.T. , 31 Dece=ber 1975, 2-5 77 Christoffersson, L. A., Lacoss, R. T., and Chinnery, M. A., Esti=stion of network =agnitude and station detection parameters, in_ Seismic Discrimination, Semi-Annual Technical Su~nry, Lincoln Laboratory, M.I.T., 31 Decenber 1975, 6-8.
~
- 78. North, R. G., and Chinnery, M. A., =b esttnation for large events in the PDE catalog, i_n_Seis=ic Discrimination, Semi-Annual Technical S - ary, Lincoln Laboratory, M.I.T., 31 Dece=ber 1975, 11-12.
79 Chinnery, M. A. , Da= aging earthquake probability studies in the Eastern U.S. and their potential application to nuclear power plant siting (abstract), Geol. Eoc. A=. Abstracts with Progrt=s, 8,, 150-1, 1976.
- 80. Lacoss, R. T., Chinnery, M. A., and Christoffersson, L. A., Ele =ents of a new statistical approach to magnitude esti=ation (abstract),
EDS Trans. Am. Geophys. Union, 5],, 285, 1976.
- 81. Chinnery, M. A., Lacoss, R.- T., and Christoffersson, L. A., Esti=ation o f =b by a network of U.S. seisnic stations (abstract), EOS Trans. A=. Geophys. Union, 57,, 285, 1976.
- 82. Landers, T. E. and Chinnery, M. A. , Spectral analysis of earthquake time series (abstract), Esrthcuake Notes, LT, 17-16, 1976.
- 83. Chinnery, M. A. (Editor), Seismic Discrimination, 25th Semi-Annual Technical Su==ary, Lincoln Laboratory, M.I.T., 30 June 1976.
8L. Chinnery, M. A. , and North, R. c. , Comparison of recent esticates of magnitude bias, M Seismic Discrimination, Semi-Annual Technical S"-m /, Lincoln Laboratory, M.I.T., 30 June 1976, 7-9 ,
- 85. Lacoss, R. T., and Chinnery, M. A., Seismic =agnitude simulation studies, in_ Seis=ic Discriminat. ion, Semi-Annual Technical Su==ary, Lincoln Laboratory, M.I.T., 30 June 1976, 9-13
- 86. Chinnery, M. A., and Lacoss, R. T., Magnitude differences berveen station pairs, M Seis=ic Discrimination, Semi-Annual Technical Su==ary, Lincoln Laboratory, M.I.T., 30 June 1976, 1h-15 87, Chinnery, M. A. , Numerical simulation of the =agnitude capability of a seismic network (abstract), EOS Trans. Am. Geophys. Union, 18.,hh3,1977
- 88. Chinnery, M. A., (Editor), Seis=ic Discrimination, 26th Semi-Annual Technical Sn-n7, Lincoln Labcratory, M.I.T. , 31 March 1977 89 Chinnery, M. A., and Johnston, J. C., Estiestion of station detection characteristics frem bulletin data, in Seismic Discrimination, Semi-Annual Technical Su==ar/, Lincoln Laboratory, M.I.T.,
31 March 1977, hk-6.
- 90. Chinnery, M. A., C0:puter simulation of network performance, h Seismic Discrimination, Semi-Annual Technical Su==ary, Lincoln Laboratory, M.I.T., 31 March 1977, L6-7.
- 91. Cninnery, M. A., Correlation of seismic activity with changes in the rate of rotation of the earth, h Seismic Discrimination, Ge=i-Annual Technical Su==ar*/, Lincoln Laboratory, M.I.T. ,
31 March 1977, 99
- 92. Chinnery, M. A., Major problems of geodynamics, Paper presented to NASA Earth Dynamics Su==er Workshcp, Boulder, Colorado, July 18-23, 1977
- 93. *Chinnery, M. A., Measurement of a with a global network, Tectonethysics , M,139-1Lk, S973.
9h. Chinner/, ". A. , (Editor), Seisnic Discrimination, 27th Semi-Annual Technical 5"-" ry, Lincoln Laborater/, y.I.T., 30 September 1977 95 Chinnery, M. A., and Johnston, J. C. Saturation of ab #0"I in Seismic Discrimination, Set:.-A".nual Technical Su urf, Lincoln Laboratory, M.I.T., 30 September 1977, 32-33
,e--. - , ,.- - - , - , - -
_ _ _ , --y,- -
- 96. Chinnery, M. A., (Editor), Seismic.Discricinatico, 28th Sesiannual Technical Su=:ary, Lincoln Laboratory, M.I.T,, 31 March 1978.
- 97. Chinnery, M. A. , A study of naxicum possible earthquakes, Annual Report, URC Contract IIRC-OL-77-019, Lincoln Laboratory, M.I.T. ,
15 August 1978 (reproduced as :TEC Publication NUREG/CR-0563).
- 93. Chinnery, M. A., (Editor), Seismic Discrimination, 29th Se=iannual Technical Su= mary, Lincoln Laboratory;- M.I.T. , 30 September 1978.
99 Johnston, J. C., and Chinnery, M. A., Measurement of md using SRO data, in, Seismic Discrimination, Sesiannual Technical Su= mary, Lincoln Laboratory, M.I.T., 30 Septe=ber 1978, L6-7 100. Chinnery, M. A., (Editor), Seismic Discricination, 30th Semiannual Technical Su==ary, Lincoln Laboratory, M.I.T., 31 March 1979 101. Chinnery, M. A. , Towards the elimination of bias in body vave magnitude, paper presented to the USOS Conference on Earth-quake Parameters, Denver, CC, March 19-21, 1979 102. Chinnery, M. A., Scatter in observed =b values, in Seismic Discri-mination, Smiannual Technical Su= ar/, Lincoln Laboratory, M.I.T., 3 March 1979, L8-9. 103.*Chinnery, M. A. , A comparision of .the seisticity of three regions of the Eastern U.S., Bull. Seisn. Soc. g., 69_, 757-772, 1979 10h. Chinnery, M. A., Seismic requirements fer a seismic data center, paper presented to US/ USSR /U Cc prehensive Test Ban Treaty Technical Delegations, M.I.T., August 10, 1979 105. Chinnery, M. A., (Editor), Seic=ic Discri=ination 31st Se=iannual Technical Su==ary, Lincoln Laboratory, M.I.T., 30 Septe=ber 1979 106. Gann, A. G. and Chinnery, M. A., Seismic data center design, in Seismic Lascrimination, Semiannual Technical Su==ary, Lincoln Laboratory, M.I.T., 30 Septe ber 1979, 1-2. 1CT. Chinnery, M. A. (Editer), Design of a Sei::ic Data Center, special internal report to DARPA, 230pp. , 30 Septe ber 1979 103.*Chinnery, M. A. and Gann, A. G., Advancei data canagement and proceccing techniques for seismic applications, paper presented at the International Symposiu: cn Meiern Computerized Methods of Registration and Interpretation of Seismic Ob:ervations, Yalta, USSR, October 2k-30, 1979 To be published in conference proceedings. 3 - - -
, u -. --- ,. .g.----
109. Chinnery, M. A. (Editor), Seistic Discrimination, 32nd Semiannual Technical Summary, Lincoln Laboratory, M.I.T., 31 March 1980. 110. Chinnery, M. A. and Cann, A. G., Background and current status, 13 Seismic Discrimination, Semiannual Technical Summary, Lincoln Laboratory, M.I.T. , 31 March 1980. 111. Chinnery, M. A., Generation of a synthetic arrivals list, J.n,_nSeismic Discrimination, Semiannual Technical S"- av, Lincoln Laboratory, M.I.T., 31 March 1980. 112. Chinnery, M. A. and Tiberio, M. A., Investigation of the properties of CD Network III using synthetic data, U. S. Delegation paper US/GSE/7 presented to the Cocmittee on Disarmament, Geneva, 7 July 1980. Also presented to NATO Advanced Study Institute on Identification of Seismic Sources, Oslo, Norway, 8-18 Septem-ber, 1980. 113. Chinnery, M. A. and Gann, A. G., Design and development of a seismic data center, U. S. Delegation paper US/GSE/6 presented to the Committee on Disarmament, Geneva, 7 July 1980. Also pre-sented to NATO Advanced Study Institute on Identification of Seismic Sources, Oslo, Noriay, 8-18 Septe=ber 1980. e Gb ~ , -w -
-sva n- - g - -r-- .-}}