Embed Size (px)
Citation preview
V a. O O ÜJ
CO LÜ
I SI)- IH-70-149
ESD ACCES§JONLIST ESTI Call No.
Copy No. of cys.
ESD RECORD COPY RETURN TO
NTiriC & TECHNICAL INFORMATION DIVISION (fSTI). BIMI.DING !
Technical Note 1970-15
Signal Processing Results
for Continental Aperture
Seismic Array
Jack Capon
15 May 1970
Prepared for the Advanced Research Pro
under Electronic Systems Division Contract A V 19(628)-5167 by
Lincoln Laboratory MASSACHUSETTS INSTITUTE OF TECHNOLOGY
Lexington, Massachusetts
AD">07*
This document has been approved for public release and sale; its distribution is unlimited.
MASSACHUSETTS INSTITUTE OF TECHNOLOGY
LINCOLN LABORATORY
SIGNAL PROCESSING RESULTS FOR CONTINENTAL APERTURE SEISMIC ARRAY
JACK CAPON
Group 22
TECHNICAL NOTE 1970-15
15 MAY 1970
This document has been approved for public release and sale; its distribution is unlimited.
LEXINGTON MASSACHUSETTS
The work reported in this document was performed at Lincoln Laboratory, a center for research operated by Massachusetts Institute of Technology. This research is a part of Project Vela Uniform, which is sponsored by the Advanced Research Projects Agency of the Department of Defense under Air Force Contract AF 19(628)-5167 (ARPA Order 512).
This report may be reproduced to satisfy needs of U.S. Government agencies.
ll
ABSTRACT
The processing of short-period P-wave data from a continental aperture
seismic array is considered. The array consists of sites located at the Large Aper-
ture Seismic Array (LASA) in eastern Montana and Long Range Seismic Measurement
(LRSM) stations located in North America. In particular, the feasibility of recogniz-
ing the arrival of the pP phase, making use of P-pP differences in velocity across
such a large array, is considered. In addition, the determination of the P-wave source
structure of an event is considered by using the array to essentially steer many beams
in the vicinity of the epicenter of the event. The capability of the array to perform
these two functions is evaluated and discussed in detail.
Accepted for the Air Force Franklin C. Hudson Chief, Lincoln Laboratory Office
iii
I. INTRODUCTION
It has been found, using data from the Large Aperture Seismic Array (LASA)
in eastern Montana, that valuable discriminants can be based on the use of the spectral
ratio for the P-wave, and the relation between surface-wave and body-wave magni-
tude, M -m, . A recent experiment has determined the effectiveness of the LASA
to discriminate between earthquakes and underground nuclear explosions using these
two, as well as other, discriminants based on results obtained from a large popula-
tion of events. However, these two discriminants are not useful at LASA below a
certain magnitude threshold. It is not clear that the spectral ratio defined for LASA
data will provide effective discrimination results when obtained at other stations.
This statement is based on preliminary data from sites other than LASA. It is pos-
sible to use the M -m, discriminant effectively at other stations. However, the S D
utility of this discriminant is limited by the difficulty in detecting the Rayleigh sur- 2
face wave due to the poor signal-to-noise ratio on the long-period instruments, as
compared to that on the short-period sensors. Thus, in attempting to lower the identi-
fication threshold by using world-wide observations, it is desirable to utilize dis-
criminants other than those based on the spectral ratio and M -m,.
One effective way of lowering the identification level is to use discriminants
based solely on the characteristics of the P-wave, due to the high signal-to-noise
ratio on the short-period seismometer. The purpose of the present work is to in-
vestigate such discriminants using data from LASA, and Long Range Seismic Mea-
surement (LRSM) sites located in North America. In particular, a depth discriminant
based on the detection of pP will be considered. In addition, the P-wave source struc-
ture of the event will also be considered. In order to accomplish both of these objec-
tives, it will be necessary to perform delay-and-sum processing, or beamforming,
of short-period P-wave data. These data are obtained from sites located within an
aperture of continental dimension, usually about 4000 km.
An extensive investigation has been made of beamforming of short-period 3
P-wave data at LASA. It was found that the performance of the beamforming process
was degraded due to the distortions introduced in the P-waves at LASA because of
inhomogeneities in crustal structure at both the receiver and the source. It is also
possible that inhomogeneities in the earth's mantle may exist which contribute to the
distortion of the P-wave. It would be expected that these effects would be even more
severe for P-waves observed over a continental aperture which is much larger than
that of LASA, i. e. , 4000 km as contrasted with 200 km. This was indeed found to
be the case, so that beamforming over continental apertures does not appear to yield
meaningful results. This will be discussed in greater detail subsequently.
II. DESCRIPTION OF CONTINENTAL ARRAY
The continental array considered consisted of sensors located at LRSM sites
in North America and at LASA. In reality, the subarray straight sums at LASA were
employed since this has the effect of improving the signal-to-noise ratio of the short-
period P-wave data. The site designations and coordinates are given in Table I. A
map which shows the locations of the sites is presented in Fig. 1.
The response of the short-period seismograph system which is typical of
that used at the LRSM stations is shown in Fig. 2. This response corresponds to
that of either a large or small Benioff short-period seismometer. The response of
the short-period seismometers used at LASA is slightly different than that given in
Fig. 2, and in the subsequent analysis will be considered to be the same as that for
the LRSM sites. The calibration data for the sensors at both LASA and the LRSM
sites were available so that it was possible to weight the data so as to maintain a
uniform calibration level across the continental aperture seismic array.
The data from the LRSM sites was available only on FM analog magnetic
tape. Thus, it was necessary to digitize the data employing an analog-to-digital
converter, using 14 bit quantization with one bit for the sign of the amplitude of
the data. These digitized data were then merged with the LASA data which were al-
ready available in a digital format on magnetic tape. The resultant tape was then in
a format which could be read directly on the IBM 360/65 using available computer pro-
grams.
TABLE I
LIST OF STATIONS SITE SITE DESIGNATION SITE COORDINATES
LATITUDE LONGITUDE PEG. D
NORTH) M S
PEG. WESr
D M 0
S AX2AL Alexander City, Alabama 32 46 38 86 07 48
BEFL Belleview, Florida 28 54 19 82 03 52
FKCO Franktown, Colorado 39 35 12 104 27 42
HL2ID Hailey, Idaho 43 33 40 114 25 08
HNME Houlton, Maine 46 09 43 67 59 09
KCMO Kansas City, Missouri 39 21 21 94 40 17
KNUT Kanab, Utah 37 01 22 112 49 39
LAAO LAS A, Montana 46 41 19 106 13 17
LAF1 LAS A, Montana 47 22 17 105 11 12
LAF2 LASA, Montana 45 54 35 105 29 14
LAF3 LASA, Montana 45 58 21 107 04 45
LAF4 LASA, Montana 47 24 37 106 56 53
LCNM Las Cruces, New Mexico 32 24 08 106 35 58
MNNV Mina, Nevada 38 26 10 118 08 53
MOID Mountain Home, Idaho 43 04 19 116 15 56
NPNT Mould Bay, Northwest Territories 76 15 08 119 22 18
PGBC Prince George, British Columbia 53 59 50 122 31 23
SV3QB Schefferville, Quebec 54 48 39 66 45 00
RKON Red Lake, Ontario 50 50 20 93 40 20
WH2YK Whitehorse, Yukon Territory 60 41 41 134 58 02
III. DISCUSSION OF COMPUTATION METHODS
We now describe the methods which were used to determine depth and P-wave
source structure using data from a continental aperture seismic array. The data
from each sensor in the array were prefiltered with a bandpass filter whose frequency
response is shown in Fig. 3a. The corresponding impulse response of the filter is
given in Fig. 3b and the response of the filter to a step of sine wave of frequency 1 Hz
is shown in Fig. 3c. The data were sampled at a frequency of 20 Hz and the wave-
form used at each site was 102. 4 seconds long, yielding 2048 samples. The fast
Fourier transform was employed to transform the data sequence into the frequency
domain and the frequency response shown in Fig. 3a was applied directly in this domain.
The filtered data were then obtained by transforming the result. The 3-db response
frequencies of the filter are 0. 64 Hz and 2. 3 Hz, which corresponds to the frequency
band in which most of the P-wave energy lies. It should be noted that the filter is
phaseless in the sense that no phase distortion is introduced. This particular filter
was chosen because it produced very little distortion in the data. There is a small
precursor which is introduced as seen from Fig. 3c. However, an analyst usually
aligns the traces so that this precursor is not included in the processing of the data.
In Fig. 3c we see that very little distortion is introduced after about one second from
the onset time of the sine wave.
The purpose of the continental aperture seismic array computer program is
to determine the depth of an event by recognizing the arrival of the pP phase, making
use of P-pP differences in velocity, and to determine the P-wave source structure of
the event by essentially steering many beams in the vicinity of the epicenter of the event.
The data consist of the digitized and merged records from LASA and LRSM sites
located in North America, as mentioned previously.
The program operates by assuming that the epicenter, but not the depth, of
the event is known, as well as the P-wave arrival times at all the stations in
the continental array. These times are computed from travel time tables, using the
known station locations and the event epicenter. The P-waves at the various stations
are brought into time alignment by an analyst. All beamforming operations are done
on the aligned traces and the time delays used are computed from travel time tables,
relative to the previously computed P-wave arrival times. This procedure is used
in the hope that the need for using station corrections is eliminated. Amplitude
weights are employed and are determined by using the calibration data for the stations,
and by compensating for the different P-wave attenuation caused by the different dis-
tances of the stations to the epicenter.
The presence of the pP phase is determined by performing a beamforming
operation over the network stations at successive times and with proper interstation
delays corresponding to trial depths ranging from 10 to 200 km. The power of the
beam output over the appropriate 1-second interval is computed as a function of the
trial depth and the trial depth that corresponds to the peak of these powers is then
taken as an estimate for the true depth.
The computed, or possibly known, value of depth is used in the determination
of the P-wave structure of the source. A beamforming operation over successive
2-second intervals, starting with the P-arrival time, is performed. This results
in a plot of power vs various latitudes and longitudes about the source hypocenter
at the computed depth, with one contour plot for each 2-second interval. A theore-
tical network beam pattern is also computed and plotted in a similar manner to pro-
vide a basis for comparison with the experimental results.
Several experiments were performed to determine the validity of the results
obtained using the computer program. These results will now be described. One of
the experiments dealt with the effect of power level changes as a function of time, for
the seismic data. In this experiment an exponentially damped sinusoid, whose fre-
quency was 1 Hz, was generated at a number of stations. The stations used were the
same as those to be discussed subsequently in the processing of the 27 October 1966
Novaya Zemlya event. The waveforms at all of the stations were made to be identical,
indicating that the data were being generated by a stationary source located at a fixed
point on the earth which was taken to be the epicenter for the previously-mentioned
event. Thus, when the beamforming process is performed over successive two-
second intervals of data, the location of the peak of the output beam power should re-
main stationary at the assumed epicenter.
However, it was found that the location of this peak did not remain stationary,
but shifted by a considerable amount. This shift of position was due to power level
variations along the waveform. In other words, it is possible that, when two-second
segments of the waveform at each sensor are shifted using time delays corresponding
to a source location other than the epicenter, the decrease in beam power output due
to the misalignment of the traces is more than compensated for by the increase in
power level in each of the waveforms. This effect, due to power level variations,
is undesirable and should be removed, since otherwise there would be very little
credence placed in the source structure computation.
In order to remove this effect, compensation for power level variations was
employed in the beamforming process. This was accomplished by measuring beam
power output as
M /N V » = T I T, W. .D. .^ ) 1 k-1 Vj=l J'1 J'k+tji/
where
M W2 . = T D2 _
J»i 11 jfk+t..
and D. , is the data in the j sensor at the k time sample, t.. is the time delay re-
quired in the j sensor for the i source location, N is the number of sensors and
M is the number of data samples contained within the integration time of two seconds,
which for 20 samples per second would be equal to 40. In the previous beamforming
operation where no power compensation was used, the preceding formula applies with
W. . = 1. In essence, this weighted beamforming operation is measuring the coherence J»1
of the delayed waveforms and is relatively immune to power level variations. When
this beamforming operation was applied to the artifically generated data mentioned
previously, the location of the peak power output remained stationary as was desired.
Another experiment was performed to test the validity of the P-wave
source structure computation using the weighted beamforming operation which
compensates for power level variations. In this experiment four source locations
were assumed at the points X, A, B, C, as shown in Fig. 4. The point X corresponded
to the location of the epicenter for the 27 October 1966 Novaya Zemlya event, to be
discussed subsequently. A pulse is assumed to propagate from a given source
location (point X, A, B, or C), to each sensor in the array, which is the
same as that discussed previously, with a travel time determined
by the time required for a compressional wave to propagate from point X to the given
source location plus the time required for a P-wave to travel from the given source
location to the sensor. This latter time is determined from travel time tables by using
the distance from the given source location to the sensor. The time required for a
compressional wave to propagate from point X to a given source location is equal to
the distance between these points, in km, divided by 6 km/sec. The amplitudes of
the pulses from X, A, B, C were assumed to be 1, v2/2, 1/2, sFl/4, respectively,
corresponding to successive 3-dB increments in power level. This sequence of four
pulses at each sensor was applied to the bandpass filter described previously to make
the data look somewhat like actual observed seismic data.
The results of the source structure computation are shown in Fig. 4, which
shows contours of -10 log (P./P ), where P is the maximum value of P.. The 0 l max' max I
level of the contours varies from 0 to 3 dB in increments of 1 dB. A grid of ±1° in
distance is given about the epicenter, located at the point marked X, and the power is
computed at 21 x 21, or 441, points. The results are presented only for four two-
second intervals, or a total of eight seconds, following the onset time of the P-wave.
The dotted contour surrounding the epicenter, marked X, in Fig. 4 indicates the locus of
points of possible sources of scattered P-waves. The results shown in Fig. 4 indicate
that the location of the peak power output shifts by an amount which is in reasonably
close agreement with the known shift in location of the source of the P-waves. That
is, the location of the zero in this figure shifts successively from point X to A, to B,
and then to C. Thus, a reasonably accurate computation of source structure has been
obtained.
A more stringent test of the method was performed by assuming a shift of
source location in two opposite directions simultaneously, instead of a shift in a
single direction as was done previously. In this case, the shift in the location of the
peak output power was unable to follow the shift in location of the source. Thus, it
would appear that the source structure computation is useful only in those situations
where the source location is moving in a single direction. The weighted beamforming
operation was also employed in the depth computation, in a manner similar to that
used in the P-wave source structure computation.
It was found that there were very significant differences in waveshape at
each station. These differences are believed to be caused by the complex crustal
structure at each site, as well as at the source. An attempt was made to design
equalization filters which would compensate for the distortion in waveshape introduced
by crustal structure. However, these attempts produced results which were believed
to be largely unsuccessful and thus will not be presented. Hence, it appears to be
impractical, at the present time, to remove the effects introduced by crustal in-
homogeneities.
10
IV. RESULTS FOR DEPTH AND P-WAVE SOURCE STRUCTURE COMPUTATION
We now present the results obtained by using the continental aperture seismic
array computer program to process the data from eight events. The parameters for
the events considered are given in Table II. These events consist of three presumed
underground nuclear explosions from Novaya Zemlya, one presumed underground
nuclear explosion from Eastern Kazakh and four earthquakes. The earthquakes were
chosen on the basis that at least four stations, not necessarily any of those used in the
processing of the event, reported the presence of a pP phase foT it to the U. S. C. G. S.
The epicenters for the eight events are also located on the map in Fig. 1.
The bandpass filtered waveforms, as well as the results of the P-wave source
structure computation, for the eight events are displayed in Figs. 5 through 20. It is
also possible to determine the sites which were employed, for a particular event, from
these figures. These sites are displayed in order of increasing distance from the epi-
center. The display of the results for the P-wave source structure computation is simi-
lar to that in Fig. 4. A typical example of a theoretical network beam pattern at 1 Hz
is shown in Fig. 21. This beam pattern applies to the collection of sensors which was
used in the processing of the 27 October 1966 Novaya Zemlya event, and shows that
typically a resolution of about 0. 2°, or about 20 km, is achieved with a continental
aperture seismic array. The beam pattern is obtained by computing the power at the
i-th position, corresponding to a given latitude and longitude in Fig. 21, as
2 P. =
l
1 N i2nt
N j=l
where t.. and N were defined previously. The power in Fig. 21 is displayed in db,
11
TABLE II
LIST OF EVENTS
K3
EVENT NO.
DATE REGION ORIGIN TIME (GMT)
LAT. (DEC)
LONG. (DEC)
DEPTH (km)
(CASA)
DEPTH (km)
(USCGS)
mb
(USCGS)
1 27 Oct 66 Novaya Zemlya
05:57:58 73. 44N 54.75 E - 0 6.3
2 18 Dec 66 Eastern Kazakh SSR
04:57:58 49. 93N 77. 73E 0 5.9
3 9 Feb 67 Colombia 15:24:47 2.85N 74. 89W 106 58 6.3
4 14 May 67 Chile- Bolivia Border Region
08:38:33 20.56S 68. 92W 114 109 5.2
26 Sep 67 Near Coast 16:11:24 of Central Chile
30.04S 71.53W
6 15 Oct 67 Near Coast of Nicaragua
08:00:50 11.86N 86. 02W
7 21 Oct 67 Novaya Zemlya
04:59:58 73. 37N 54.81E
8 7 Nov 68 Novaya Zemlya
10:02:05 73.41N 54.86E
22
80
55
162
0
5.7
6.2
5.9
6.3
relative to the maximum power of unity, in steps of 1 db from 0 to 6 db.
The results of the depth computation provided by the continental aperture
seismic array (CASA) program are given in Table II, along with the depths reported
by the U. S. C. G. S. The corresponding positions in time are denoted on the waveforms
presented in Figs. 9, 11, 13, and 15. It is seen that in only one of the four cases con-
sidered was there agreement between the CASA program depth and the U. S. C. G. S.
depth. That is, for event number 4 there is good agreement between the two depths.
However, for this event the presence of the pP phase is quite clear on the individual
records, cf. Fig. 11. In the other three cases the presence of the pP phase is not
clear on the individual traces and the CASA program detected the presence of this
phase at a time which appears to yield a depth significantly different than that reported
by the U. S. C. G. S. This happened in spite of the fact that there were four stations on
the earth that reported the presence of this phase to the U. S. C. G. S. Thus, it would
appear that the determination of depth by detecting the presence of the pP phase with
a continental aperture seismic array is relatively unreliable. This is probably due
to the effects of the crustal inhomogeneities mentioned previously, as well as other
effects which may not as yet be well understood.
We now consider the results for the P-wave source structure computation, and
begin by considering event number 1, which is the 27 October 1966 Novaya Zemlya
event. The waveforms from the various sites which were used are shown in Fig. 5.
The stations are distributed over an aperture of about 4300 km and vary in distance
from the epicenter from 46° to 74°. For this event, as well as the other presumed
underground nuclear explosions, the determination of depth by the CASA program is
not meaningful and the depth is assumed to be zero km in the P-wave source structure
13
computation.
The waveforms of Fig. 5 show striking differences in waveshape for this event
as viewed across the network. For example, the waveforms observed at RKON,
HNME are very simple while those at LASA, i. e., the subarray straight sums from
LAAO, LAF1, LAF2, LAF3, LAF4, are extremely complex. These data show that
some caution must be exercised in employing the complexity criterion for discrimina-
tion.
The results of the P-wave source structure computation for event number 1
are shown in Fig. 6 (a-d). The X in each figure, as well as those figures for the
other events, represents the location of the epicenter provided by U. S. C. G. S. and
the zero corresponds to the location from which the peak power output is provided by
the weighted beamforming process. In Fig. 6a the zero and X coincide, indicating
that during the first two seconds after the origin time the peak power is coming from
the epicenter. In Figs. 6 (b-d) we see a migration of the location providing the peak
power. These locations are indicated in Fig. 22 on a map of the Novaya Zemlya region.
The migration of the location of the zeros in Figs. 6 (b-d) is compatible with
conditions imposed by known seismic compressional wave velocity in the crust. That
is, the locations of the zeros lie inside the dotted contour surrounding the epicenter
marked X. This dotted contour was described previously in connection with Fig. 4
and is the locus of points of possible sources of scattered P-waves during the time
interval defined by the particular frame. If a point marked zero should be obtained
outside of this contour then it is known immediately that an erroneous result has been
obtained. Thus, some credence may be placed in the P-wave source structure com-
putation for event number 1. However, the results for two other Novaya Zemlya
14
events, numbers 7 and 8, shown in Figs. 18 and 20, respectively, do not yield consis-
tent results since in many cases the location of the zero falls outside the dotted con-
tour. This also happens quite often with the other events, cf. Figs. 8, 10, 12, 14,
16. Thus, it appears that the determination of the P-wave source structure by means
of beamforming of continental aperture seismic array data is relatively unreliable.
Once again, this is probably due to the effects of both source and receiver crustal
inhomogeneities which were mentioned previously, as well as other effects.
15
V. CONCLUSIONS
The feasibility of detecting the arrival of the pP phase has been considered,
using weighted beamforming of data from a continental aperture seismic array. It
was found that this procedure was relatively unreliable, probably because of the
effects of crustal, and possibly mantle, inhomogeneities. This conclusion is based
on only a small population of events consisting of four earthquakes. However, it
appears that the conclusion is valid since it agrees with similar conclusions obtained
4 by Chiburis and Ahner on the basis of a large population of events. Thus, those
previous reports which indicated promising results for processing of data from con-
tinental aperture seismic arrays to detect the presence of the pP phase should be
considered as unduly optimistic.
It was also found that the determination of the P-wave source structure using
weighted beamforming of data from a continental aperture seismic array was un-
reliable, once again due to the effects of crustal and mantle inhomogeneities. The
theory of plate tectonics ' holds some promise for understanding these inhomo-
geneities and possibly compensating for them. If this were possible then there might
be some hope in obtaining meaningful results from the processing of data from sensors
located over a continental aperture. In the lack of such an understanding, the beam-
forming of data from a continental aperture seismic array must be considered as
yielding unreliable, if not meaningless, results.
16
REFERENCES
1. R. T. Lacoss, "A Large-Population LASA Discrimination Experiment," Technical
Note 1969-24, Lincoln Laboratory, M.I.T. (8 April 1970), DDC AD-687478.
2. J. Capon, "Investigation of Long-Period Noise at the Large Aperture Seismic
Array," J. Geophys. Res. , 74 , 15 June 1969.
3. J. Capon, R. J. Greenfield, R. J. Kolker, and R. T. Lacoss, "Short-Period
Signal Processing Results for the Large Aperture Seismic Array," Geophysics,
33, June 1968.
4. E. F. Chiburis and R. O Ahner,"The Comparative Detectability of pP at LASA,
TFSO, UBSO, and CPSO," Seismic Data Laboratory Report No. 231, 18 April
1969.
5. D. P. McKenzie, "Speculations on the Consequences and Causes of Plate Motions,"
Geophys. J. R. Astr. Soc. , jj, September 1969.
6. D. Davies and D. P. McKenzie, "Seismic Travel-Time Residuals and Plates,"
Geophys. J. R. Astr. Soc. , JJS, September 1969.
17
150 180
cc
150 120
Fig. 1. Locations of stations and events.
18-2-9332
1.0
PERIOD (sec)
Fig. 2. Frequency response of the short-period seismometer system.
19
to O
'•0r
o a
a. |o2
0 0.5 2.0 2.5 3.0 FREQUENCY (Hi)
(a)
II-2-»II
I i i i i I i i i i t i i i i I -5 0 5 10
(b)
3.5 4.0 '5 -5 1 0
1 1 1 1
TIME (sec)
(c)
1 • 5
1 10
Fig. 3. Frequency response, impulse response and stepped sine wave response of bandpass filter.
c 3
1\ o—
118- 2 - 9059- 21
c
1J i \ \\ ^ 0~ LTj \ \\ \ ' Iw) >) \
('v^x ) \ /
\ /
(
(a) 0 to 2 SEC (b) 2 to 4 SEC
i
54 56
LONGITUDE (East)
(c) 4 to 6 SEC
58 52
'/\U\.
54 56
LONGITUDE (Eo«t)
(d) 6 to 8 SEC
Fig. 4. P-wave source structure result for artificially-generated seismic data.
21
S,TE l"-6»^-?l
SVQB "* ■' —» ■■ ■ — »*^|\M^^»V»-V\/\r^»^rv>^^^^ v>^V>*»*-^>»AA«»»*»»<*^» ■»« WNA.w^^vv*.-»»^^Ay^ y^w»,^^.. «.. ^... ■%„ ,.M<>V%» , »y»,.v. «^ ^
PGBC • ■■ ■ ^-»AM/Wi/W^r^AwV^^V^^V^^i^Ar^/V-'*^^ *rt^» ■w^»»*N»»1'-" » »VW»'»» .^^»^-^w»y» ^v»«.*^VVV-/»^»WV«~i^>»W»<^^^%wv—
RKON ... ■■■ ■ - ^MM/^^J\fS^^ ^^^.»^wM^^tv^^^^^yAWM^^^ -. .. ..-■ . »v-<v
LAF1 — ■ - ■■J^W^/\^r^^^^«»r-s^^A»^»^>'s — -^ —^.■'«^^■■^»^■o^w^^.. ..HV ■^r».>v^^.Vt-»-v^»W- ■»■ ..y»» w ■»» ^ .. . ■ ^-, »v
LAF4 ■ - -■ J\f*$\fftfr+J\f\J\[W^
l_AAO ^WV\A^A/V>AA^,>*<A^^^AA^AAA^^A' ^A^-<^*^<^^<'^^V^V»VV^^^<^"^''^'^^'^V^'>V*,*^^^'W^^^ ~"" "" -
KCMO N^WWV-I
MNNV - ^»»—» — >»-«A»'^^*»«*w^»»-^^'w. ■ —^r^ ■- .■»»».■««vwxv^i »•
KNUT - «»ArfV>^p.^^^/WV'V^A^^'>^W'*^*^'*^,VV**>A'l*^>*^>* iw^^fVV^'Vi'^^VN'^V»''^«''^''1 .^.*w»<w^^r«>"'»-v-
AXAL ' ^|V\/V«^M^VVv*>v^v<w^''vvV,s,%' ■»vv^ ■^«'■■»^^^^«»■1 ^^o^»»wi»— ^ - <^i.
BEFl - . .■.»■>...,/^A»<y^w»....,.v»^w.'»v>'WtfV>-^w»>'
10 20 30 40
TIME (sec)
Fig. 5. Waveforms for 27 October 1966 Novaya Zemlya event.
22
l)8-?-9?89l
^
3
TOI]V)
I..
f^
V ^>
# <> x 3
a 3^2
(a) 0 to 2 SEC (b) 2 fo 4 SEC
a -3
72 5
p< -3
\
\)
K5
/ / v "IJ *
\ \
< x /
/ / /
^"
£ (?
54
LONGITUDE (Eo$t)
54
LONGITUDE (Ea«t)
(c) 4 to 6 SEC (d) 6 to 8 SEC
Fig. 6. P-wave source structure result for 27 October 1966 Novaya Zemlya event.
23
SITE
NPNT-
1 18-2-9290 1
-J. ffi'^MMWv^^\AA/v^A^^ •^*/**ft*tf**v-++^~-++. ■ VV^/Sr«^^lA-■■■**■■»■-^»^V-w—^.-.. ^» -. ,,v»..».yvV»'wv^-.v^*»V
WHYK- -vv- ■<W>- ■ .w^»^»^« -.w-»v» «—. A ■* .• ■*+. ..-^w^-
P6BC
RKON- -fftt*" w*-^«.»-*»
HNME- -^W"
LAF4 -J—\MA\/WW^^A/WtJ\^ «^-.vV^^VWvA/VMAVW^ .^VWW>»V»W*A
ho 4^
LAF 1- VViN/lNHHwv^^ **M\/*\/\f\r*dKrJ\/~>r**A/>J\fV^^ VAr WVAA/V"VW^-«>V>rvS^>«> . ^v«-^ -Vo->
LAAO- ■»/| /fo<VW\As/VM)'/V**'«VW•»«¥*- Al/~~*»t-'*~<+*~~ -^»■v^w- ■■■- ■■■v«.«*»«v-
LAF3- 'jkt^{TJ\J\r**t***'v*i*+u*-'"</\r*-miS',+' »v ■*»• wv/wvJ\pv>-" ——»*«*—■ ■ ■««*■—^- ■ ^—^. y^ ^> v><^. v«vw
LAF2- ■^^A^^/SA/>^^*UMA'A^~*\/^/^/\/^^*^>''^^^>*V^ •••^VV^V«>W««VW\A^-^^-V^^WS>»«(Ar^ *^V^'»lVV -A,«,-»..»^.....^—•■*>>«•■
MOID "AiVyV^,^,'^>"^^"*vv- -vw i^wAWi—-ß*+*^m*^ -*-«*^,-w—.v *>»\/WVA*-»**-.v-W "*^ .*..A/>—«V- ■"^■vV »»*/».«■
MNNV- -^ ■A^«>A^,«^VV^^^ ■ ^■•»*»*v-
KNUT- ■n/\pr^^AArVV^»"»^A<v-t/\/V»r-/»- A^ANv. ■»^ ^-v*.^--^«——-*^«.^.
to 30 40 SO
TIME (sec) 9C
Fig. 7. Waveforms for 18 December 1966 Eastern Kazakh event.
ll-?-9?9l
i ( x
(o ) 0 to 2 SEC (b) 2 to 4 SEC
!E ̂
~ 50 c^
^ ¥
O •.*£
0
77 78
LONGITUDE (Eo»t)
Ä X
'1
<fc3-
7»
LONGITUDE (Eott)
(c) 4 to 6 SEC (d) 6 to 8 SEC
Fig. 8. P-wave source structure result for 18 December 1966 Eastern Kazakh event.
25
SITE
HNME-~
FKCO —
KNUT,
18-2-9292
'T/Y^WlfWY*1*^^ Vwtf^ >v -*/W\ft *** ■ wy» W^^«V*»'tW»^^^^««*V>»«
.N*^-*vAA/WV«»*>*r>/W»vWVv,»--,»^^>*-*«"- ^-^AAA'S^SA'»^ *»«*-^*^«^ v^^^^JUA^^yW^ ^N^A^^A^M^/CA^^^^^^-VA^A^^
.■s. »■vvwv-'V>»vV\/«v\A/' ■^«VyVvV»«»^- .•^■%' ̂ ^MrW\^/W\A/ VW\f^ «^ •m***AAf\f\J\/ V --^A/*» wv/wyv^
^VrA.^^^r ■* ■■ v»-V\MA""V' ^\T'J\f>r*4^>^\f\f>^ - ■* *-* ^V^- J\f*fr**-mJ\f* RKON-
to
MNNV-
MO40
PGBC
«•»•^■■«VV '»^V^—\/V»,'*V^**l/V>Ar oV^^A>^r^l^*V*^*^^^■^^^*^-V/^/V^*^/>l^^V/^^*,***^^^^*^ ■/VN-»v^v»
WHYK- IU*^^^I v^».y w^^-»'»% ' i ■*
J I I L
NPNT ■
10 20 50
USCGS
40
PP CASA
TIME (see)
50 60 70 60 90
Fig. 9. Waveforms for 9 February 1967 Colombia event.
2
3
k
B> • < if
$
«3
'<)
>
<•}
^
(o) 0 to 2 SEC (b) 2 to 4 SEC
O fl
D
w 6
<5>
P7
75.3 750 745
LONGITUDE (We*t)
74.0
"xT
/ 4 6
\ i— ^
\o
Ä
T*
A V
755 75 Ö 74 5
LONGITUDE (W««t)
(c) 4 to 6 SEC (d) 6 to 8 SEC
Fig. 10. P-wave source structure result for 9 February 1967 Colombia event.
4 1
740
27
SITE
KNUT
18-2-9294
LAF2- ^f \ A/^M^^^VW'^/SA—^y\/W^^^>^Ar^W■
RKON
MNNV
H
-^JV^
\fsj\j\fv^^
yW^/VNvVV\A»Awv\j^^
WV—»«•■»»•■ *»•»»■ -«w»*^ ^ ^»v*.«.
-^/V—A/——
LAF3- ■^s • | ^yvw/>^^M^^A-*»^V- </W._VvV*W\/-v<V
LAAO
SVOB AATWV/BWW-V^ jvM^v/y*/^^^—,/v^*«.*».
11
LAF4 '
MOID
PGBC
WHYK
-^j^/W^^-v.-^ •■» ,>V>--v—■^»■»> ^IVWVWW^V-M^^^A^M^-^ w-VVnA>Af ■VWy/V*
'y\Af-**#*A ^^f*wWWV
TIME (sec)
Fig. 11. Waveforms for 14 May 1967 Chile-Bolivia Border Region event.
28
A 2 k \ \ 3
\<
\ (
1l-2-9?S5
(o) 0 to 2 SEC (b) 2 to 4 SEC
Q 3
( 2 ■lOu 7 ( [4 \
/" / / \ / \
/ /
\ \ \
1 s { ; \ / \ /
\ / / '—■>
69 5 69 0 68 5
LONGITUDE (Wt»t)
(c) 4 to 6 SEC
69 5 690 6« 5
LONGITUDE (Wt«t)
id) 6 to 8 SEC
Fig. 12. P-wave source structure result for 14 May 1967 Chile-Bolivia Border Region event.
29
O
SITE
LCNM
HNME
LAF2 /-
LAF3 —
LAAO
LAF1
HLID
LAF4
SVOB
-10
I-2-9296 I
-y^».».». ^^»^/^^A^»^^^^/^A^/\^A^^A^A"l■V"^^ <^»V <~-^/>f—<**^+v\r~-— .»I^K/W'WVWIAAA ^» .>^..»»^»^V»yv%i
*<^V»^i ."» ■ ^- ^»^v^A^tf1 I^^V^-IA^WV ^—^^. AV^^^'»',^W <^M^ «>^i
KNUT y\f^y>AAAA^^^Af^^^A^ —Ay^x\A^
i
<~*v~-s**/**>^s*f**<* ^\AA*\^/\tW\f*^\f^^\fi>^/\/^\r^^ *"• •*>*+~-*—*~***s*-*\r''*,m>S****^+-*
^/VwWl/V^^ :ffllw '^ww^y^^
10 20 30
PP CASA
PP USCGS
«0
TIME (sec)
50 GO 70 B0 90
Fig. 13. Waveforms for 26 September 1967 Near Coast of Central Chile event.
)
4fc>
^-3
/ l°>
\
> /
^—
(o) Oto2 SEC (b) 2 to 4 SEC
72.0 71.5
LONGITUDE (West)
(c) A to 6 SEC
720 71.5 71.0
LONGITUDE (West)
(d) 6 to 8 SEC
Fig. 14. P-wave source structure result for 26 September 1967 Near Coast of Central Chile event.
31
to
SITE
KNUT
HNME
LAF3-
LAAO ■
MNNV •
RKON
LAF4-
HLID
PGBC
WHYK-
NPNT
118-2-92981
-■-^A^-yvyv»-.>''v-V\^>--»-*»--^*--*^—-—
'V^/W\/yV\^v^~WW-'^---^--Vu'W^ —.-Vv---"-'>u—v—■ ■.».- ^^« v«— -*v—
V'^AAV^AWVWWV1^ WW-VW- »«*»*—«VW-...<\/^ I rv- »^»•»-v^.-.'» ■-■■«»v —»%/»■ Vu^*—-V-
-NA/VA*— AMM/WV~--U»V^^ ..^V^SWA^WW-
-W\/V-—■V/'V^-^-»v-v—v*-
. v^ - v»-—YVv^wv,^"u' --^^■v-v/>-^- ■»". ."»-i/^/v,^ - -"Si V" •««•■■-«»/»»
■ -v-"""-
TIME (sec) PP
USCGS
Fig. 15. Waveforms for 15 October 1967 Near Coast of Nicaragua event.
I 120
3
(o) 0 »o 2 SEC (b) 2 to 4 SEC
I 12(
uj Q 3
11.5
* (- 7-J
/ / / / / \
p (. X \ \ \ \ \ .
\ \ >
/ / /
9 / J
I
860
LONGITUDE (West)
(c) 4 to 6 SEC
(ftrÄ /I Im1
,»
/ / /
/ /
\ \ \ \
i ( X \ \ \ \
\ N
) /
/ / /
ft
865 860 LONGITUDE (West)
(d) 6 to 8 SEC
Fig. 16. P-wave source structure result for 15 October 1967 Near Coast of Nicaragua event.
33
GO 4^
SUE , [18-2-9300
NPNT u'A1/l|lrf^^/vV,-•^^ ^^^w *■»**»» A .■■«•- ~ ■ w —~-v — .—-
WHYK ■
RKON
LAF3-
HLID
^^Ml'^*'>^l^^^^*^^ ^■»•■V'VVV»"»- ■■*—~-~ -*»*»•■ - VW-—- +r-
HNME »Y^W*
LAAO "-f^—V^WW^V^^V^>A^.V--^A/W\^ ■.-^A/^V'VNA^V A^^VW^ .^-. ■—W/V •"- ]
LAF 2 -* ^»W;'^"^''-V»*'^,uV>« ••-«^ A> VV» .-«.^/V, .<V .^...^.^. .N.«^. .- -—"^)V,^-v
,1 V, M
KNUT *■—.N^^V-^V
LCNM VW/w*»**-—
20 30 40 50 60 70
TIME (sec)
Fig. 17. Waveforms for 21 October 1967 Novaya Zemlya event.
e <
&
(o) 0 to 2 SEC
J
h X i \
\ / *k
£>
11-2-1301
\
IAI
(b) 2 to 4 SEC
€4 ^s
W/1 w ^L
4 <<.
4; <3Ü
54 56
LONGITUDE (Eost)
U ü ^T
3 AT K^
0
<d
54
LONGITUDE (Eost)
(d) 6 to 8 SEC
&
<>
(c) 4 to 6 SEC
Fig. 18. P-wave source structure result for 21 October 1967 Novaya Zemlya event.
35
-;
I et
2
— 2
2 ?
ed
N
ed
> C 2 DC
o
I E > 0 2
o CO
s
> cd
ON -—I
be
36
1)1-2-9S0? 1
A'
1 3
\
M>
(a) 0 to 2 SEC (b) 2 to 4 SEC
«^ <• <^
^-3^
t$6 i
r 5 733
V
1
J /
a ►-
i X
> \ / / 6 J h-JJ
1
\
V
54 56
LONGITUDE (East)
V
q
s^a>-
.a.
tJ
A
ii&
3j<2
&
t a
54 56
LONGITUDE (Eoit)
'* r3
(c) 4 to 6 SEC (d) 6 to 8 SEC
Fig. 20. P-wave source structure result for 7 November 1968 Novaya Zemlya event.
37
52 54 56
LONGITUDE (East)
Fig. 21. Typical theoretical network beam pattern at 1 Hz.
38
Q 3
74 5
sr" /^k * ^\ 3
KRESTOVAYA GUBA^"^. N/ ^\ /
74 0
BARENTS SEA
V^v^ ^/ GORA PERVOUSMOTRENNAYA ^\ J
73 5 J PIK SEDOVA \
V /^N GORA KUPLETSKOGO / O J J^tv r\ /
^Ji[ GORA ^^^^y] Vs/ -—~^\J>i MOISEYEVA N_ V, J
/POLUOSTROV \\ MATOCHKIN PAN KOVA SHAR
ZEMLYA "*■ ^^->
'S 73 0
o / , SV^f^V-JMYS SHUBERTA
•^2fv> PUKHOVOYE )
to 1
oo
7? 5 «1 54 56
LONGITUDE (East)
Fig. 22. Migration of location of peak power for 27 October 1966 Novaya Zemlya event.
39
UNCLASSIFIED
Security Classification
DOCUMENT CONTROL DATA - R&D (Security classification of title, body of abstract and indexing annotation must be entered when the overall report ia classified)
1. ORIGINATING ACTIVITY (Corporate author)
Lincoln Laboratory, M.I.T.
2a. REPORT SECURITY CLASSIFICATION
Unclassified 2b. GROUP
None
3. REPORT TITLE
Signal Processing Results for Continental Aperture Seismic Array
4. DESCRIPTIVE NOTES (Type of report and inclusive dates)
Technical Note 5. AUTHOR(S) (Last name, first name, initial)
Capon, Jack
6. REPORT DATE
15 May 1970
7a. TOTAL NO. OF PAGES
44
7b. NO. OF REFS
6
8a. CONTRACT OR GRANT NO. AF 19(628)~5167
b. PROJECT NO. ARPA Order 512
9a. ORIGINATOR'S REPORT NUMBER(S)
Technical Note 1970-15
9b. OTHER REPORT NO(S) (Any other numbers that may be assigned this report)
ESD-TR-70-149
10. AVAILABILITY/LIMITATION NOTICES
This document has been approved for public release and sale; its distribution is unlimited.
11. SUPPLEMENTARY NOTES
None
12. SPONSORING MILITARY ACTIVITY
Advanced Research Projects Agency, Department of Defense
13. ABSTRACT
The processing of short-period P-wave data from a continental aperture seismic array is considered. The array consists of sites located at the Large Aperture Seismic Array (LASA) in eastern Montana and Long Range Seismic Measurement (LRSM) stations located in North America. In particular, the feasibility of recognizing the arrival of the pP phase, making use of P-pP differences in velocity across such a large array, is considered. In addition, the determination of the P-wave source structure of an event is con- sidered by using the array to essentially steer many beams in the vicinity of the epicenter of the event. The capability of the array to perform these two functions is evaluated and discussed in detail.
14. KEY WORDS
seismic discrimination LASA LRSM
P-wave data beamforming
underground explosions signal processing
plf-1800 40 UNCLASSIFIED
Security Classification
