ML20195C822
| ML20195C822 | |
| Person / Time | |
|---|---|
| Issue date: | 05/23/1986 |
| From: | Zurflueh E NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| To: | Reiter L Office of Nuclear Reactor Regulation |
| References | |
| NUDOCS 8605300618 | |
| Download: ML20195C822 (10) | |
Text
-
MAY 2 31986 MEMORANDUM FOR: Leon Reiter Senior Reliability and Risk Anal' cst Reliability and Risk Assessment Branch Division of Safety Review and Oversig.t. NRR FROM:
Ernst G. Zurflueh, Geophysicist Seismology Section Earth Sciences Branch Division of Radiation Programs &
Earth Sciences, RES
SUBJECT:
EVALUATION OF PAPER BY 20BACK ET AL.
The enclosed paper by Zoback et al., which we received from you, describes an analysis of repeated triangulation measurements in the vicinity of New York.
The measurements indicate high rates of crustal strain on Western Long Island and in the Ramapo area. These high strain rates may give an indication of greater seismic hazard potential in these areas than has been assumed to date.
It is, therefore, important to analyze the validity of the conclusions drawn in the paper.
In the course of recent phone conversations with Dr. W. B. Strange of the National Geodetic Survey (NGS), I have found out that the NGS has evaluated this paper using the NGS data base which is larger than that used in preparing the paper. The NGS analyzed both the Long Island and Ramapo areas.
On Long Island, the majority of the anomalous strain rates seem to be derived from stationc located on sand bars or other unstable ground and, therefore, may not be real. After reaching this preliminary conclusion, no further work was done in this area.
More detailed studies were made in the Ramapo area. A draft report on this work will be issued soon. First, an analysis was made using only the data used by Zoback and adhering to his methods of analysis. The results from this analysis were identical to those described in the paper, indicating that there 4
is no error in the analysis. However, a repeated analysis using the larger data base showed less evidence for an anomalous strain rate. The conclusion drawn from the study is that there may be anomalous strain in the Ramapo area, I,
but that the anomaly is not very large. Also, it has to be noted that the
~
Zoback study was based mostly on data measured between 1938 and 1956.
In summary, it appears that the Ramapo area does have a strain rate that is higher 4
than normal although not to the extent postulated in the Zoback paper.
1 I
1.
8605300618 860523 PDR MISC 8605300618 PDR
-.. ~ -
A.
Leon Reiter 2
MAY Z 31986 The question of what the exact strain rate is and how it relates to recent seismicity needs to be resolved. With this in mind, a meeting was held in December 1985 between NGS personnel and two of the authors of the paper (Zoback and Prescott). At the meeting it was decided that, if funding is available, approximately 40 stations in the Ramapo area will be remeasured. Funding may be made available by the Empire State Electric Energy Research Corporation (ESEERCO), RES will monitor the development on this topic and keep you in-formed as progress warrants.
Ernst G. Zurflueh, Geophysicist Seismology Section Earth Sciences Branch Division of Radiation Programs &
Earth Sciences, RES
Enclosure:
Paper by Zoback et al.
Distribution /R-2811:
(DCS/PDR RMinogue DN: beratan/ll Efrc/Chron Dross LBeratan ESB Sbj/Rd KGoller AMurphy EZurflueh n --
bYb_s "[l RIJ)~ -
ISub,j ect Filo T0.'
. _ __ 7 y..,.,,
- 7. :,,
g y;
- -- j I
77 _
y
}
ESB:RES:pf ES R ESB:RESgf
.~
_. c EZurflueh AMurphy LBeratan
'y'y l
73 f g/86 y/,{.3/86
-j,.. m
'T I
F /23/86 f
Eft
O Macminin Joumals Ltd,1985 i
Evidence for lower crustal ductile 1969; 31 angles in Eastern Long Island,27 angles in Western gg gy Long Island,26 angles near the eastern and western boundaries of New Jersey, and 15 angles in eastern Pennsylvania.
He deviatoric part of the strain field can be described by two Mark D. Zoback*, William H. Prescott average strain rate parameters, is,92, which define shear strain
& Scot W. Krueger*
rates on NW-SE and E-W striking planes, respectively. He
, change in an observed angle over a particular time period can US Geological Survey, Menlo Park, Califomia 94025, USA be written as Aa = at-[(sin 20 -sin 28.)f,+(cos 20 -cos 28 )f2]
2 2
Historic triangulation data haie been analysed to determine where Aa is the observed change in angle, at is the time between whether intraplate seismicity is associated with ongoing ductile the two observations, and e, and # are the azimuths of the two 2
deformation in the lower crust.De model we have attempted to sides of the angle'2. i, and i2 were estimated from the observed test is basically analogous to strain accumulation and release angle changes by a least-squares process". An equivalent way of resenting the strain field is in terms of the azimuth of the P
along plate-boundary strike-slip faults like the San Andreas Fault P ane across which shear is maximum l
in California. That is, beneath an elastic-seismogenic upper crust
~20 km thick, strain is preferentially localized within ductile shear j tan ~'( fi/ f2) + 90*,
if f, < 0 and i3> 0 zones in the lower crust due to broad-scale plate driving forces.
9" tan.i(i,/ f2) otherwise The localized lower-crustal ductile strain causes stress and strain to accumulate elastically in the brittle crust which is eventually and the shear strain rate across that plane released in crustal earthquakes. At greater depths, this localized j, g g3 shear deformation probably develops into pervasive ductile flow.
Numerous geodetic measurements along the San Andreas Fault As shown in Table I and Fig.1,the groups of angles in Pennsyl.
confirm that earthquakes in the brittle upper crust are produced vania, New Jersey, and Eastern Long Island showed no statisti.
by the release of clastic strain that results from ongoing ductile cally significant strain rate at the 95% (2a) confidence level.
shear or slip in the lower crust' 2. We have found evidence of high However, the Western Long Island and New York-Connecticut rates of crustal deformation in sosthern New York which suggest groups of angles both show statistically significant strain, with that localized ductile shear is occurring in the lower crust.
maximum strain occurring on NNW-SSE and N-S striking i
ne presence of ductile shear zones in the lower crust has planes, respectively.
been suggested by workers who have studied geological struc-In the Western Long Island and New York-Connecticut data tures within exhumed shear zones'dand effects of temperature sets, the strain rates are surprisingly quite large, roughl7 and pressure on the dominant deformation mechanisms of crus-equivalent to strain rates observed near the San Andreas fault.
tal rocks'. Laboratory studies suggest several mechanisms that he critical question is: how realistic are the estimates of stan-may be responsible for the development of such zones in lower dard deviations quoted in Table I and Fig. I because if the crustal conditions". Several authors have suggested that.sig-quoted standard deviations are correct, then it is unlikely that nificant deformation may occur along discrete large-scale shear the observed strains result from errors.He original observations y;
zones within continents , and play a key role in continental of most of the angles used in this study were third order surveys.
fragmentation". In the context of the model tested here, these nese angles have errors that are normally distributed with a shear zones would be sites of localized ductile shear in the standard deviation of 1.2 s (ref.14). De later surveys were intraplate lower crust.
mostly first or second order with standard deviations of 0.6-0.8 s
>j We have analysed repeated triangulation measurements made (ref.14). The strain. rate parameters were calculated from a j
by the National Geodetic Survey (NGS) in three areas: the weighted least squares adjustment. The weights used for the region of the Ramapo fault zone of northeastern New Jersey observations were obtained from the NGS data file and varied if and south-east New York,the New Madrid area of southeastern from 2.78 s72 to 0.6 s-2 corresponding to standard errors in the Missouri and northeastern Arkansas, and the Charleston, South observations of 0.6 s to 1.3 s. The standard deviations given in Carolina area. Large earthquakes occurred near Charleston in.
Table I and Fig. 2 were obtained by propagating the error 1886 and near New Madrid in 1811-12. All three areas are estimates of the observations thrcugh the matrix equation used
, ~
currently seismically active.
to calculate 9, and 92. De error estimates are not dependent To establish horizontal control points for map makers, on the fit of the observations to the time.and. space-uniform engineers and others, NGS routinely measures the angles in model. He a postchori standard deviation of an observation of networks of control points. He primary purpose of making the a priori unit weight does depend on the fit of the observations measurements is not the detection of crustal motion and the to the model. This quantity is angles are not systematically reobserved. However, the demands
-#hr
[\\
for additional control often require that areas be resurveyed and in the process some angles are reobserved. Only the shear strain a.
/
)
- 8" components of the strain tensor can be determined from repeated
. nm.
i angle observations.
{
While insufficient repeated angles were available to obtain where o, is the observed ith angle, c, is the calculated ith angle, useful results in the New Madrid and Charleston zones,130 a, is the a priori standard error of the ith angle, and n,t is the a
repeated angle measurements were available in the region of number of degrees of freedom.
the Ramapo fault zone. For the purpose of analysis, we split If the observed modelis a correct description of the deforma.
these angles into five separate groups (Fig.1). One group of tion and if the a priori standard deviations accurately describe angles in New York and Connecticut is near the northeastern the scatter in the observations then a. will have a value of 1.
end of the broad zone of contemporary seismicity associated For the five data sets in Table I, a,is 1.4,0.7,1.3,1.3 and 1.2.
with the Ramapo fault system. nese angles were measured at We conclude that the a priori error distribution of the observa-various times between 1872 and 1973. Outside the seismicity tions is reasonable and consequently that the standard deviations zone: 100 repeated angles were measured between 1885 and for the strain components are reasonable.
To test further that the computed strain rates in the New 8
9 85 " *
- emeni.44=,er nep.nm.e or ce.pny c.. s4 r a u r
- y. s-r a. c.i.r m2m. usr m on Dep.m em et c i.sy, u.=,wy or c.w.,=
a,ewey, to undetected measurement errors or blunders m the data, each car.,= m20 us4 <s w K4 group of angles was split into two independent subsets. In both
- e
76*
rs' y;ge 73s 77,
/
MA 42, 1
O
,ge iso.te % Eos bY A2 s N3E % 7 HMITFoRD 1
i i
9 e i.
M Fig.I Area of study in the New York.
Connecticut-New Jersey-Pennsylvania area.
NEw -.
voRx The box indicates the zone of contemporary p
seismicity associated with the Ramapo fault
+
rone. The angles analysed i's this study were S a E21 *^ & H Med inm N N gmps hn W6 an NJ i s 0. 8 2 % 0. t 3 G
A2 Nset w13 described in the text and summarized in Table
~
^Z'uNREsoWED 0
- 1. The symbols and rates indicate the orienta.
""5^"
tion (degrees clockwise from north) the magni.
a o* -peon, a e A2 s N25W % 9 lude (y strain / year) of masimum right. lateral
,o, shear strain rate.
~
Region of recent sesmic%
9 o.or % o.is
/
c A2 UNAtsoLVED p
kAngles used in 4
this study
=
U o
so tookm a
MD 39*
0E 35*
j rs' re*
73*
7*
New York-Connecticut and Western Long Island, the subsets While ductile straining in the lower crust could be quite yielded strain rates quite similar to that of the group a*s a whole complex these data indicate evidence of high rates of strain at (Table 2). Consequently, although the results indicate surprising depth regardless of any specific model. If a simple strike-slip
'arge strain rates, we can find no basis for rejecting them; they model is correct, the planes of maximum shear strain strike f
ppear to be real.
north to north. northeast whereas the mapped faults and other The cause of the anomalous strain rates observed in the New surficial peological structures in the area generally strike more f
York-Connecticut and Western Long Island areas is not clear.
casterly'. This would imply either that the surficial features are Although there is widespread low-magnitude seismicity in this not directly related to the zone of ductile strain localization at area (Fig. 2), no carthquakes of sufficient magnitude have occur-depth or that the actual deformation at depth is much mon red during historic time which coald account for the observed complicated than represented by such a simple model. In the a
h strain as coseismic deformation". Th us, the most likely explana-latter case, the true nature of strain at depth may be obscured tion of the observed strain rates is in terms oflocalized ductile by our ability to estimate only the shear components of the 2
deformation in the lowe'r crust.
strain field.
p The most common model used to interpret observed strain Several lines of evidence may support the contention that e
near the San Andreas Faultis one in which a vertical strike-slip lower crustal aseismic deformation could be cccurring along fault is locked to some depth, D, and slipping at a constant rate, N-S to NNW-SSE planes in this area. The NE-SW direction b, at depth greater than D. Then the shear strain rate of the of maximum horizontal compression that is implied by the shear j,
surface at a distance x from the fault trace is given by deformation is quite reasonable for this area. It is consistent g
with the direction. defined by numerous stress data in western g
g New York"". Recent stress measurements within the area f(x)= D [IH1/DH 2
covered by the New York-Connecticut data
- indicate NE-SW 2
If such a model describes the deformation reported here, deep-compression (Fip' 2), as does a reanalysis of focal mechanisms crustal aseismic slip would be occurring in the lower crust at in New England'. Also, a N-S striking trend of seismicity runs rates of several centimetres per year. Such rates could probably through the area of the New York-Connecticut data" and recent not be sustained in the lower crust for more than several hundred microcarthquakes near the Hudson river ~30 km north of New years without a major earthquake occurring.
York City suggest that carthquakes in the area align along N-S Table I Strain rate parameters on fne groups separated according to the number of angles used in the analysis No. of Dates of i,
12 i
Arimuth Area angles observations (prad yr)
(prad yr~')
(prad yr)
t New York-Connecticut 30 1862-1973
-0.02 + / - 0.04
-0.18 + / -0.05 0.18 + / -0.05 N3* E + / -7 Western Long Island 27 1932-1967 '
O.43 + / -0.15
-0.37 + / - 0.15 0.57 + / - 0.14 N25' W + / - 9 Eastern Long Island 31 1939-1967
-0.19 + / - 0.11 0.09 + / -0.09 0.21 * / - 0.11 N58' E + / - 13 Northern New Jersey 26 1931-1962
-0.05 + /-0.16 0.04 + / - 0.14 0.07 + / - 0.13 N68' E + / -74 Eastern Pennsylvania 15 1885-1969
-0.06 + / - 0.20 0.10 + / -0.18 0.12 + / - 0.13 N74* E + / - 56 y,, right. lateral shear strain rate on a NW-SE striking plane, uncertainty is I s.d. fa, Right. lateral shear strain rate on an E W striking plane ~'
f, Maximum shear strain rate. Arimuth, orientation of plane of maximum shear measured clockwise from the north.
'e6 3
ys.
74*
73' ii Table 2 Suain raies of subsets of groups identi6ed in Table 8
$ 3.o 3 a l
0 2.oam No. of Deses of 9 2.s29 e 202.4 l
Subset angles observations y
9 New York-Connecticut 12 1862-1973 019 + / - 0.05 N2' i + / - 8
-)
s ieso-New York-Connecticut 18 1932-1966 017 + / -0 09 N7"I.+ / - 15
[
'\\,,
West t.ang Island 8
I864-1948 0 62 + / - 0.23 N22" W + / - 1I 8
/
West ten 5 sland 19 1932-l%7 0 60 + / - 0.19 NI8* W + / - II I
8, b
o+4 a
/ [
- [.
[ 8, '
are observing a relatively shon-term strain episode.
One of the most striking aspects of intraplate seismic zones p
g -[qq:f:hp is their apparent association with zones of ancient tecton-v s
[p +,
S
- e_-
e i co' ism'" 2*-'.Ris association has led to the widely-held presump-jp
- -5 J555fri tion that large intraplate earthquakes tend to occur in zones of f
q+
c5f 59g I crustal weaknus (that is, zones of anomalously low-strength in p'
[5f -:29 /
the seismogenic, upper 20 km, of the brittle crust). Here are h-5_f;~~DM[:g' 5 _ p-several problems with this hypothesis, however. First, no adequate physical explanation has been put forth to explain
- 5555:1-9_g-9--
J '" p which specific fault zones in the brittle crust would be expected 8
- so 3o' to have low strength. Laboratory studies show that, with the 55555:255:__g -
- 5:QU555?'
exception of several clay minerals not stable at great depth, the frictional strength of nearly all rocks and minerals are about
- 5:59:"
1
- 39 -
~. -
the same s. Available in Situ stress measurements supports the 29 P
relevance of such data for active faults. Second,in the areas near intraplate seisrr.ic zones there is no evidence of the
. i Fig. 2 Generalized geological map" showing location of micro-anomalous stress field that would be expected if the seismicity earthquakes recorded by the Lamont-Doherty - Geological were due to stress concentrations around ' weak' intrusive rocks l
Observatory from 1974 to 1981 (ref. 34). The arrows indicate the of low clastic stiffness ** In fact, the opposite seems true-direction of maximum horizontal compression determined from intraplate earthquakes seem to occur in response to a broad, recent in setu measurements in the Hudson Highlands
regionally uniform stress field Finally, it is not clear how
' weak' zones in the upper crust would be capable of accumulat-l.,
striking planes in response to NE-SW to ENE-SWS compress.
ing sufficient clastic strain energy to produce major earthquakes ion". The largest historic earthquake in this area was a witfely.
or how, once released, strain energy would reaccumulate in such J
felt event with an estimated Modified Mercalliintensity of VII zones rapidly enough to explain observed recurrence times on
~I in 1884.
the order of hundreds to thousands of years.. Because it is 32 The most surprising aspect of these results is the extremely not possible to answer such fundamental questions, the
- zone high strain rate observed. If aseismic displacements were occur.
of crustal weakness
- hypothesis has been essentially useless in j
ring in the lower crust they would have to be of the order of distinguishing which ancient tectonic zones might be expected
'i several centimetres per year to explain the data. Rese rates to produce large damaging earthquakes irrthe future.The most seem quite fast for an intraplate area. As derived from studies important characteristic of the model presented here is that sites of global plate motionathe North American Plate is estimated of large intraplate earthquakes are controlled by zones of local-l to be moving only ~2.7 cm yr-' in the vicinity of the Ramapo ized strain in the lower crust which concentrate stress clastically fault zone. It is not clear how so much internal deformation in the upper crust. Identification of these zones might indicate f
i could occur within the plate for extended geological time sites where large intraplate carthquakes could occur in the j
periods. Thus, one possibility is that, in a geological sense, we future.
Received 5 March, accepted 6 Augesa 8985
- 17. Ratct Ae, N M en FicM Sanees sa Nr. Jersey-Geologr esid Guide se FeeM Tnes 278 iRwegets
- 1. savage.J. c. A Burford, R o 1 geepe s Aes78, 832 (1973).
University, he= ark 1980L r
2 savage. J C. A Aer Eerre $ ate 5ts II. II (1983L 18 Zoback, M L & Zoback, M. D J geophys Aes85. 6113 11980s 3 Grecou J J geel See Lead 133, 257 (1977) 19 Hackman, s H. Healy. J H a Zoback, M. D 1 geophes Aes 90. $497 t19sil 4 sibson, R H J geel Sec Lead 133,19111977L
- 20. Zoback, M D., Anderson, R N A Moos. D fos 64,363 Il9sSt S %%ie, s. H. Serroms, s E carreras J.. sham N. D A Hamphreys. F. J J stamer. Gael 21 Gephardt, J W & Forsyth, D. D Geology 13, 70 (198 $ t 2,175 t19 sop
- 22. seborowsks. K D. wdhams, C. Kelleher. J = A sianon. C. T. ad seism See Aan 72,
- 6. Raictifle, R M SwA geef Soc Am 32,123 gl973L 1601 (1982).
- 7. sobson. R H J geef Soc Lead B40. 748 (l983).
- 23. Minnaer, J B & Jordon. T. H. I geophes Aes 83. 3331 t19?s e 8 Karby, s H Ace Geopers Spera Mrs 21.1438 (1983b 24 McKensee, D P. Geophys J R eier Ser 30.109 c19721 9 Moanar. P & Tapponier. P. Sticore ist. 419 (1973)
- 23. RaicliRe, N M Se.R geel Sec. Aas 82, 129 11973)
- 10. Tapponeer, P & Molnar. P. Netwee264, 389 (1976).
26 Jackson J A Notere283, 343 (19:01 II. sykes, L R Aes Geope s Space Phys 56, 621 (19781
- 27. Zobuk, M D er el Science 209, 971 11980s r
- 32. Frank, F. C. Sd seism Ser Aan Se,35 (1966t
- 28. Byerlee.3 D Ne App ( Geophys 116,615 (1978t
- 13. Prescott.W H Sd sense Sor Aan GA1847 (1974K 29 Zoback. M D & Healy, J H Aaels Geophrs 2. 689 419841 14 Federal Geodriis Comirol Commitiee Classefrenea $senderd of Arrwrery and Generet
- 30. Kane. M F. en U.S geet Sws Prof Asp No 1028 (1978)
SperVicenons of Geodent Centrof Seeters (NoAA 1980).
- 31. Zoback. M D a Zeback. M L 3ricere 21197 I1981L IS Aggarmal, Y. P & sykes, L IL Srwere 2ee,425 (1978)
- 32. Ress. D P ad geof Ser Amt90.108 3 (1979t 16 savage.J C en DtWarensas se Sehds ed Nabarro, F R. N 265(North. Holland, Amsterdam.
- 33. obermeie. s er el Scariere227.408 t1985t 1980t
- 34. Kafka. A L NUREGICR 3079(Us Nuclear Regulatory Commesuon,1983 g s g' W
e--
Leon Reiter 2
MAY 2 31986 The question of what the exact strain rate is and how it relates to recent seismicity needs to be resolved. With this in mind, a meeting was held in December 1985 between NGS personnel and two of the authors of the paper (Zoback and Prescott). At the meeting it was decided that, if funding is available, approximately 40 stations in the Ramapo area will be remeasured. Funding may be made available by the Empire State Electric Energy Research Corporation (ESEERCO). RES will monitor the development on this topic and keep you in-formed as progress warrants.
Ernst G. Zurflueh, Geophysicist Seismology Section Earth Sciences Branch Division of Radiation Programs &
Earth Sciences, RES
Enclosure:
Paper by Zoback et al.
Distribution /R-2811:
DCS/fDR)
RMinogue DN: beratan/11 Circ /Ch'ron Dross LBeratan ESB Sbj/Rd KGoller AMurphy EZurflueh ESB:RES:pf ES R
'ESB:RESgh(f j
EZurflueh AMurphy LBeratan 5 /23/86 yfg86 g/43/86
- 2) fit 4
w--
,n--
f MAY 2 31986 I
i
~
MEMORANDUM FOR: Leon Reiter 1
Senior Reliability and Risk Analyst l
Reliability and Risk Assessment Branch Division of Safety Review and Oversight, NRR FROM:
Ernst G. Zurflueh, Geophysicist t
Seismology Section Earth Sciences Branch Division of Radiation Programs &
Earth Sciences, RES
SUBJECT:
EVALUATION OF PAPER BY Z0BACK ET AL.
The enclosed paper by Zoback et al., which we received from you, describes an analysis of repeated triangulation measurements in the vicinity of New York.
The measurements indicate high. rates of crustal strain on Western Long Island j
and in the Ramapo area. These high strain rates may give an indication of j
greater seismic hazard potential in these areas than has been assumed to date.
It is, therefore, important to analyze the validity of the conclusions drawn in the paper.
In the course of recent phone conversations with Dr. W. B. Strange of the National Geodetic Survey (NGS), I have found out that the NGS has evaluated this paper using the NGS data base which is larger than that used in preparing the paper. The NGS analyzed both the Long Island and Ramapo areas.
i On Long Island, the majority of the anomalous strain rates seem to be derived from stations located on sand bars or other unstable ground and, therefore, may not be real. After reaching this preliminary conclusion, no further work was done in this area.
More detailed studies were made in the Ramapo area. A draft report on this work will be issued soon. First, an analysis was made using only the data used by Zoback and adhering to his methods of analysis. The results from this analysis were identical to those described in the paper, indicating that there is no error in the analysis. However, a repeated analysis using the larger i
data base showed less evidence for an anomalous strain rate. The conclusion i
drawn from the study is that there may be anomalous strain in the Ramapo area, but that the anomaly is not very large. Also, it has to be noted that the Zoback study was based mostly on data measured between 1938 and 1956.
In summary, it appears that the Ramapo area does have a strain rate that is higher than normal although not to the extent postulated in the Zoback paper.
1 1
~~ L _. _. -
O Macmilla Joumals Ltd,1985 Evidence for lower crustal ductile 1969; 31 angles in Eastern Lsong Island,27 angles in Western Stra.in locallZal.IOR In Southern New York Long Island,26 angles near the eastern and western boundaries or New Jersey, and 15 angles in eastern Pennsylvania.
He deviatoric part of the strain field can be described by two Mark D. Zoback', William H. Prescott average strain rate parameters, ii, ya, which define shear strain
& Scot W. Krueger rates on NW-SE and E-W striking planes, respectively. He o
, change in an observed angle over a particular time period can US Geological Survey, Menlo Park, Cahfornia 94025, USA be written as da - Ar-[(sin 20 -sin 28 )fi+(cos 20 -cos 2ei)f2]
2 2
Historic triangulation data have been analysed to deterinine where Aa is the observed change in angle, Ae is the time between whether intraplate seismicity is associated with ongoing ductile tpe two obsenati and e and 8 are the azimuths of the two n
2 2 I deformation in the lower crust. He model we have attempted to sides of the angle. y, and y2 were estima,t,ed from the observed test is basically analogous to strain accumulation and release angle changes by a least. squares process. An equivalent way along plate-boundary strike-slip faults like ate San Andreas Feult of presenting the strain field is m terms of the azimuth of the in California.That is, beneath an elastic-seismogenic upper crust plane across which shear is maximum
~20 km thick, strain is preferentially localized within ductile shear T " )1 tan-'(fi/i2) + 90*,
if y, < 0 and f2 > 0 zones in the lower crust due to broad-scale plate driving forces.
tand(fi/ f2) otherwise De localized lower-crustal ductile strain causes stress and strain to accumulate elastically in the brittle crust which is eventually and the shear strain rate across that plane released in crustal earthquakes. At greater depths, this localized shear deformation probably develops into pervasive ductile flow.
9, g gut Numerous geodetic measurements along the San Andreas Fault As shown in Table 1 and Fig.1, the groups of angles in Pennsyl.
confirm that earthquakes in the brittle upper crust are produced vania, New Jersey, and Eastern Lsong Island showed no statisti-a by the release of elastic strain that results from ongoing ductile cally significant strain rate at the 95% (2a) confidence level.
shear or slip in the lower crustb2 We have found evidence of high However, the Western Long Island and New York-Connecticut rates of crustal deformation in southern New York which suggest groups of angles both show statistically significant strain, with that localized ductile shear is occurring in the lower crust.
meximum strain occurring on NNW-SSE and N-S striking i
The presence of ductile shear zones in the lower crust has planes, respectively.
been suggested by workers who have studied geological struc-In the Western Long Island and New York-Connecticut data tures within exhumed shear zones'-* and effects of temperature sets, the strain rates are surprisingly quite large, roughl7 and pressure on the dorr.inant deformation mechanisms of crus-equivalent to strain rates observed near the San Andreas fault.
tal rocks'. Laboratory studies suggest several mechanisms that De critical question is: how realistic are the estimates of stan-may be responsible for the development of such zones in lower dard deviations quoted in Table I and Fig. I because if the crustal conditions'. Several authors have suggested that.aig-quoted standard deviations are correct, then it is unlikely that j
nificant deformation may occur along discrete large-scale shear the observed strains result from errors.The original observations zones within continents ", and play a key role in continental of most of the angles used in this study were third order surveys.
fragmentation". In the context of the model tested here, these These angles have errors that are normally distributed with a shear zones would be sites of localized ductile shear in the standard deviation of 1.2 s (ref.14). The later surveys were intraplate lower crust.
mostly first or second order with standard deviations of 0.6-0.8 s
.j We have analysed repeated triangulation measurements made (ref.14). He strain. rate parameters were calculated from a 1
by the National Geodetic Survey (NGS) in three areas: the weighted least squares adjustment. De weights used for the region of the Ramapo fault zone of northeastern New Jersey observations were obtained from the NGS data file and varied I
and south-east New York,the New Madrid area of southeastern from 2.78 s-2 to 0.6 s'* corresponding to standard errors in the i
Missouri and northeastern Arkansas, and the Charleston, South observations of 0.6 s to 1.3 s. The standard deviations given in Carolina area. Large earthquakes occurred near Charleston in Table 1 and Fig. 2 were obtained by propagating the error 1886 and near New Madrid in 1811-12. All three areas are estimates of the observations through the matrix equation used
, ~
currently seismically active.
to calculate y, and y2. He error estimates are not dependent To establish horizontal control points for map makers, on the fit of the observations to the time-and. space. uniform engineers and others, NGS routinely measures the angles in model. The a posteriori standard deviation of an observation of networks of control points. He primary purpose of making the a priori unit weight does depend on the fit of the observations measurements is not the detection of crustal motion and the to the model. His quantity is angles are not systematically reobserved. However, the demands
~ #hr for additional control often require that areas be resurveyed and in the process some angles are reobserved. Only the shear strain
\\
a.
/
components of the strain tensor can be determined from repeated n,,f angle observations.
While insufficient repeated ang!cs were available to obtain where o, is the observed ith angle, c, is the calculated ith angle, useful results in the New Madrid and Charleston zones,130 a, is the a priori standard error of the ith angle, and n.r is the repeated angle measurements were available in the region of number of degrees of freedom.
the Ramapo fault zone. For the purpose of analysis, we split if the observed modelis a correct description of the deforma.
these angles into five separate groups (Fig.1). One group of tion and if the a priori standard deviations accurately describe angles in New York and Connecticut is near the northeastern the scatter in the observations then a. will have a value of 1.
end of the broad zone of contemporary seismicity associated For the five data sets in Table I, a is 1.4,0.7,1.3,1.3 and 1.2.
o with the Ramapo fault system. These angles were measured at We conclude that the a priori error distribution of the observa-various times between 1872 and 1973. Outside the seismicity tions is reasonable and consequently that the standard deviations zone,100 repeated angles were measured between 1885 and for the strain components are reasonable.
To test further that the computed strain rates in the New York-Connecticut and Western Long Island areas were not due M 7[u"EmTNN=[.#N.~u=NCri cYrD.Ny'.
to undetected measurement errors or blunders in the data, each r
caww=. wao, usus.w xa.
group of angles was split into two independent subsets. In both J
's.
f e
767,
rs' ye*
ys*
yy*
4 8
+
/
y4 i
i 42' I
O
- 42' 4 s 0 ts *^ 0 05 NY 1
A2 e N3E % y MARTFoRo PA RI CT a
I l
4f
. 4 i' Fig. I Area of study in the New York-Corinecticut-New Jersey-Pennsylvania area.
ug,
vonx The boa indicates the zone of contemporary p
seismicity associated with the Ramapo fault
+
rone. The angles analysed in this study were u 012. 0. t3 NJ t s 421 % 011 divided into the 6se groups shown which are 0
A2 Nsst %is desenbed in the tent and summarised in Table A2. uwAEsooto O
- l. The symbols and rates indicate the orienta-
" ' ^
a n (degrees clockwise from north) the magni-so* -mt40ELP*+e4 e A2 s N2sw % 9 40, gyde (p strain / year) of maximum right-lateral shear strain rate.
Region of recent
[
G 38 0 07 % 0.13 A2
- uNAEsoLVEC p
Angles used in 4
ttus study t
U 0
50 100km MD 39*
- 3s' rs*
7.*
rs' 72*
Ncw York-Connecticut and Western Long Island, the s0bsets While ductile straining in the lower crust could be quite yielded strain rates quite similar to that of the group as a whole complex these data indicate evidence of high rates of strain at (Table 2). Consequently, although the results indicate surprising depth regardless of any specific model. If a simple strike. slip
'arge strain rates, we can find no basis for rejecting them; they model is correct, the planes of rraximum shear strain strike north to north-northeast whereas the mapped faults and other ppear to be real.
ne cause of the anomalous strain rates observed in the New surficial peological structures in the area generally strike more Q
York-Connecticut and Western Long Island areas is not clear.
casterly'. nis would imply either that the surficial features are Although there is widespread low. magnitude seismicity in this not directly related to the zone of ductile strain localization at area (Fig. 2), no earthquakes of sufficient magnitude have occur-depth or that the actual deformation at depth is much more t
red during historic time which could account for the observed complicated than represented by such a simple model. In the strain as coseismic deformation".Rus,the most likely explana-latter case, the true nature of strain at depth may be obscured
.g
- X tion of the observed strain rates is in terms of locali2ed ductile by our ability to estimate only the shear components of the
'E deformation in the lowe'r crust.
strain field.
p He most common model used to interpret observed strain Several lines of evidence may support the contention that
-e near the San Andreas Fault'* is one in which a vertical strike. slip lower crustal aseismic deformation could be occurring along fault is locked to some depth, D, and slipping at a constant rate, N-S to NNW-SSE planes in this area. He NE-SW direction b, at depth greater than D. Then the shear strain rate of the of maximum horizontal compression that is implied by the shear surface at a distance x from the fault trace is given by deformation is quite reasonable for this area. It is consistent g
with the direction, defined by numerous stress data in western g
g New York"" Recent stress measurements within the area f(x) = -[l HxMW covered by the New York-Connecticut data ' indicate NE-SW D
2 If such a model describes the deformation reported here, deep.
compression (Fig'.2), as does a reanalysis of focal mechanisms crustal tseismic slip would be occurring in the lower crust at in New England. Also,a N-S striking trend of seismicity runs rates of several centimetres per year. Such rates could probably through the area of the New York-Cor.necticut data" and recent not be sustained in the lower crust for more than several hundred microcarthquakes near the Hudson river ~30 km north of New years without a major earthquake occurring.
York City suggest that earthquakes in the area align along N-S Table I Strain rate parameters on Eve groups separated according to the number of angles used in the analysis No. of Dates of y,
y, y
Azimuth Area angles observations (prad yr-')
(prad yr-')
(prad yr-8) t New York-Connecticut 30 1862-1973
-0.02 + / - 0.04
-0.18 + / - 0.05 0.18 + / - 0.05 N3* E + / - 7 Western Long Island 27 1937-l % 7 '
O.43 + / - 0.15
-0.37 + / - 0.15 0.57 + / -0.14 N25' W + / - 9 i
Eastern Long Island 31 1939-1967
-0.19 + / -0.11 0.09 + / -0 09 0.21 * / - 0.11 N58* E + / - 13 l
Northern New Jersey 26 1931-1962
-0.0$ + /-0.16 0.04 + / - 0.14 0 07 + / - 0.13 N68' E + / - 74
[ astern Penns>hania 15 1885-l % 9
-0.06 + / - 0.20 0.10 + / -0.18 0.12 + / - 0.13 N74* E + / - 56 y,, right. lateral shear strain rate on a NW-SE sinking plane-', uncertainty is I s.d. y,, Right lateral shear strain rate on an E-W striking plane-'.
y, Maximum shear strain rate. Azimuth, orientation of plane of mavimum shear measured clockwise from the north.
,m...
~
3
, s.
74 F3' i i Table 2 Strain rates of subsets of groups identined in Table I
$ 3038 I
sw 1
no..f o.m of e 2 s.2.
Subsel angles observations y
9 e 2or4 8
New York-Connecticut 12 1862-1973 019 + / -0 05 N2* E + / - 8
-)
4 e so-New York-Connecticut 18 1932-1966 017 + / - 0 09 N7* E + / - 15 3
g West tong Island 8
1864-1948 0 62 + / -0.23 N22'W + /-ll f
f
=,,
8 w
/
West tong Island 19 1932 1967 0 60 + / - 0.19 N!8* W + / -II b
/) /,.
o k
are observing a relatisely short-tenn strain episode, f(
e
- [,
One of the most striking aspects of intraplate seismic zones f
j 55 is their apparent association with zones of ancient tecton-v g i3[5'5[:.
- woo' ism' " 2*-2.This association has led to the widely-held presump-g
[
o
_5
- 559 I
Gt,Erv%7 tion that large intraplate earthquakes tend to occur in zones of
_-5#
35g I crustal weakness (that is, zones of anomalously low strength in
. 59 ' -2 the seismogenic, upper 20 km, of the brittle crust) There are f.
,,_j? f)59 /
_ -3 5
several problems with this hypothesis, however, First, no
% e:-~
7 * :'
adequate physical explanation has been put forth to explain d? =
which specific fault zones in the brittle crust would be expected
- 5I:-55k.::u[f5#
8
(($55$(:5(
- ao so*
to have low strength. Laboratory studies show that, with the
- 55555555'-
exception of several clay minerals not stable at great depth,the 1
frictional strength of nearly all rocks and minerals are about
- 25555 --
1) the same. Available m siru stress measurements supports the
$55-P
[
relevance of such data for active faults #'. Second,in the areas near intraplate seismic zones there is no evidence of the j
Fig. 2 Generalized geological map" showing location of micro-anomalous stress field that would be expected if the seismicity canhquakes recorded by the Lamont-Doherty Geological were due to stress concentrations around ' weak
- intrusive rocks Observatory from 1974 to 1981 (ref,34). The arrows indicate she of low clastic stiffness'* In fact, the opposite seems true-
+
direction of maximum horizontal compression determined from intraplate earthquakes seem to occur in response to a broad, recent in stru measurements in the Hudson Highlands *.
regionally uniform stress field" Finally, it is not clear how 3
i
- weak' zones in the upper crust would be capable of accumulat-striking planes in response to NE-SW to ENE-SWS compress.
ing sufficient clastic strain energy to produce major earthquakes, 32 ion. The largest historic earthquake in this area was a witfely-or how, once released, strain energy would reaccumulate in such
' ~d felt event with an estimated Modified Mercalliintensity of VII zones rapidly enough to explain observed recurrence times on l
in 1884, the order of hundreds to thousands of years'2", Because it is The most surprising aspect of these results is the extremely not possible to answer such fundamental questions, the ' zone high strain rate observed. If aseismic displacements were occur-of crustal weakness
- hypothesis has been essentially useless in 3 j ring in the lower crust they would have to be of the order of distinguishing which ancient tectonic zones might be expected
~, i several centimetres per year to explain the data. These rates to produce large damaging earthquakes irrthe future. The most j
seem quite fast for an intraplate area. As derived from studies important characteristic of the model presented here is that sites 22 of global plate motion,the North American Plate is estimated oflarge intraplate earthquakes are controlled by zones of local.
.j to be moving only ~2.7 cm yr in the vicinity of the Ramapo ized strain in the lower crust which concentrate stress elastically fault zone. It is not clear how so much internal deformation in the upper crust. Identification of these zones might indicate could occur within the plate for extended geological time sites where large intraplate earthquakes could occur in the periods. Thus, one possibility is that, in a geological sense,~ we future.
Receeved 5 March, accegned 6 Augen l985
- 17. Ratchfle N M in Faeld snedes sa htelerser Geetogread Ges.de ss Feld Taps 27B iReisers B. sa*nge, J c a Berford, R. o I geopA*s Ars 78, 832 119738 University, heeark,1980)
- 2. $ avage, J C. A Re feash pieace Sra II, it (198H, 18 zebut, M L & zebuk, M D J seapest Aes SS,etilt19 sos 3 Groccat,3 J geed Ser Lead 133,257 f1977) 19 H ckman, s H, Healy,J H & zoback M D / geophrs Act90, 5497 Iles!)
4 sebson, A H J geef Ser. Lead 133,191 t 1977).
20 zoback, M D, Anderson, R N & Mees, D EOS %36) (1985
$ whee. s H, Evere=s, s L. carreras, J, shee, N. D & Hamphreys. F. J J senere Geel
- 21. Gephardt, J w a Forsyth. D D 6eedore 13. To i19856 2,175 t1980s
- 22. seborowshe, K. D, kihams, G. Kelleher, J A & saatten, c. T. Seit seisas $ar Am 72.
6 Ratclifle, R M Sea geel Sac Am 82,12$ I1971).
1608 (19821
- 7. sibson, R H J geod Ser imod 846,74B (1981)
- 23. Minseer, J B A Jordon, T. H J g* apers Aes 83. SHI Ilt?8 P 9 Orby, s H Ace Geophys Spore Pnys II,1458 (19th 24 McKentee, D P Geopara J A. e.se 5er 30,109 419723 9 Molnar, P & Tapponier, P. Scie =re 189, 419 11975) 23 Reich 8e, N M SJ geef 5er Ane $2.12)I1971I 30 Tapponier, P & Molnar P hesere364, 189 119766 26 Jackson,3 A Nevere 283,34) tl9so)
II. sykes, L R Ars Geophra $ pere Ph s 16, 623 (1978)
- 27. zehack. M D es el Sarare300, 971 (1980)
F 12 Freak. F. C. anA saum Ser Am Se,31 (1966) 28 Byerlee,J D Phee Appi Geophrs 1945 el$ I1978) 13 Prescott, w H Sea saum See Ana te51847 t1974)
- 29. zebuk. M D & Healy, J H Anais Gespers 2, est Il9641 14 Federal Geodetic consrol committee, classultens 5sendeed e/ Arrecery and cwar si 30 Kane, M F en (15 geel San hof Php No 1025 11978)
Spen /irensas e/Geoderar ceareef Serveys iNoAA,1980)
- 31. zebuk, M D & zeback, M L 3rware 213,97 t19516.
l$ Aggarwat, Y. P & sykes, L R. Science 3ee,42511978) 32 kass D P ad geef 5er A= 90,301)(1979 16 sasage,J C.sa Duderenons m 5eads ed habarro,F. R N.241(North Holised Amsterdam, 33 Obermeer, s er el Srware227,400 t19:51 1980) 34 Kufha, A. L NUREGlcR 3079 (Us heclear Regolasery commassen,1984
,~r
- _ _ - ~. -
- - -