puvt:]

EVALUATION OF THE IA ROE ARRAY LOme-PERIOD HETWORK

Htm men stwue 4/Mir Amnu tmnn

PvpruHucvd i »

NATIONAL TECHNICAL INFORMATION SERVICE

>piin<)ii«id. v« ;7i»i

ueutun n. wi

AIH mice rtcmcAi MPHICMTIOKS CIHTM WnUW*. O.C.

■a?r haltet ¥IIA UHIFOnH

AOVAMID HCStAKH MOJICTS A6EMY Hnltar Mmittimi Rnunh 0/fiet

AHM Otto Il:1i20

^TTELEOVNE GEOTECH AUXAmiA lABORATOmS

APM0¥10 FOH NilIC KllASl OlSTOItUmH UHUMITtO.

I

**»* a.™.,., * ' '* **""* "-"•* "*" ""*. t* »„ „n, r„Me„ A„Ha,„„ c,Mf>

Tix0M E

Ilnc-ln«t«;i fißcl S<curHy CUntflotton

DOCUMENT CONTROL DATA • R&D

I OWIOIM»TIMO AcfltflfV ffiigSS «yllUfJ

TELEDYNE GEOTECII ALEXANDRIA, VIRGINIA

I RI»MT TITLf

i* «IOOMT (icuaiT* ckMiinetfieV Unclassified

Ik oaoup

EVALUATION OF THE LARGE ARRAY LONG-PERIOD NETWORK

«UTNORm n.«V( Mm«. SnlMäw! iüSSSfJ

Mack, Harry

December 17, 1971

»■ VHOJICT NO

VELA T/1709

ARPA Order No. 1620 rfARPA Program Code No. 1F10

7« TOTAL NO. Or ■>*•(•

35 76. NO or ntw»

• • OniOINATOK-l niPoatT MUMtmrSi

** SJi'Slponl*0*'' t>0(" ^n»'oU'»">""»»'» «•> may »• •••(«)•«

IP « VAILAIILITV/LIMITATION NOTICES

APPROVED FOR PUBLIC RELEASE; DISTRIBUTION UNLIMITED.

II (UPPLCMCNTARV NOTCI I» «PONtORINO MILITAHV ACTIVITV

ADVANCED RESEARCH PROJECTS AGENCY NUCLEAR MONITORING RESEARCH OFFICE WASHINGTON. D. C.

It ABSTRACT

Rav^?ain™ f^ysfs of LASA long period data has shown that the S f8^ rieteCtl0n rate is ^uite Poor for event:s located to the south of the array and to be best for events from the north- SAP •CCrpariSOnu0f KIJrile earthquakes recorded by LASA and NORSAR indicates that the Rayleigh wave detection rates are

?omi nran3r^wntS rith b0dy Wave magnitudes greater than or equal

to .,0 and that surface waves from such small events are con- sisttntly detected. Preliminary studies of central Asian and western Soviet Union events recorded by any of the three laree arrays shows that surface waves are detected from earthquakes with body wave magnitudes of 4.0 but the data set is too incomplete as yet to establish a network threshold.

Al!^-1 ariy Processing technique is introduced which allows the detection and measurement of small amplitude Rayleigh waves.

I« KEY WORDS

Large Arrtiy Network Evaluati Large Array Network

on

L Unclassified

Security Classification

HVALUATION OF THE LARGE

ARRAY LONG-PERIOD NETWORK

SEISMIC ARRAY ANALYSIS CENTER No. 4

AFTAC Project Number:

Project Title:

ARPA Order No.:

ARPA Program Code No.!

VELA T/1709

Seismic Array Analysis Center

1620

1F10

Name of Contractor: TELEDYNE GEOTEC11

Contract No.:

Effective Date of Contract:

Amount of Contract:

Contract Expiration Date:

Project Manager:

F33657-71-C-0510

1 December 1970

$1,034,946

31 December 1971

Wm. C. Dean (703) 836-7647

P. Q. Box 334. Alexandria, Virginia

APPROVED FOR PUBLIC RELEASE; DISTRIBUTION UNLIMITED.

>

This research was supported by the Advanced Research

Projects Agency, Nuclear Monitoring Research Office, under

Project VELA-UNIFORM, and accomplished under the technical

direction of the Air Force Technical Applications Center

under Contract F33657-71-C-0510.

Neither the advanced Research Projects Agency nor the Air

Force Technical Applications Center will be responsible for

information contained herein which has been supplied by other

organizations or contractors, and this document is subject to

lacer revision as may be necessary. The views and conclusions

presented are those of the authors and should not be interpreted

as necessarily representing the official policies, either expressed

or implied, of the Advanced Research Projects Agency, the Air

Force Technical Applications Center, or the U S Government.

^

ABSTRACT

A continuous analysis of LASA long period data has

shown that the Rayleigh wave detection rate is quite

poor for events located to the south of the array and

to be best for events from the northwest, A comparison

of Kurile earthquakes recorded by LASA aid WORSAR

indicates that the Rayleigh wave detection rates are

similar for events with body wave magnitudes greater

than or equal to 4.0 and that surface waves from such

small events are consistently detected. Preliminary

studies of central Asian and western Soviet Union events

recorded by any of the three large arrays shows that

surface waves are detected from earthquakes with body

wave magnitudes of 4,0 but the data set is too

incomplete as yet to establish a network threshold,

A novel array processing technique is introduced

which allows the detection and measurement of small

amplitude Rayleigh waves.

izT

TABLE OF CONTENTS

Page No.

INTRODUCTION 1

CONTINUOUS ANALYSIS OF LASA LP DATA 2

ARRAY COMPARISONS 3

LASA-NORSAR 3

ALPA-NORSAR 5

THE NETWORK 7

SIGNAL PROCESSING DEVELOPMENTS 8

^

LIST OF FIGURES

Figure Title Figure No.

Location of regions used in LASA continuous analysis. 1

Raylcigh wave detections/non-detections for region 1. 2a

Rayleigh wave detections/non-detections for region 2. 2b

Rayleigh wave detections/non-detections for region 3. 2c

Rayleigh wave detections/non-detections for region 4. 2d

Rayleigh wave detections/non-detections for region 5. 2e

Rayleigh wave detections/non-detections for region 6. 2f

Rayleigh wave detections/non-detections for region 7. 2g

Rayleigh wave detections/non-detections for region 8. 2h

SAAC mb/Ms for suite of Kurile events recorded at LASA. 3a

SAAC m^/Mg for suite of Kurile events recorded at NORSAR. 3b

Rayleigh wave detections and non-detections for Kurile suite at LASA. 4a

Rayleigh wave detections and non-detections for Kurile suite at NORSAR. 4b

Ms - mu plot for central Asian events recorded at both ALPA and NORSAR September - November 1971. 5

^

LIST OF FIGURES (Cont'd.)

figure Title Figurc No

Ms " mb Plot for central Asian events recorded at ALPA and/oi NORSAR September - November 1971.

-JC

6

Ms - m|j plots for western and southwestern Soviet Union events recorded at NORSAR. 7

NORSAR beam showing coherent noise propagating from the east, masking a Rayleigh wave from a west Russian event, 3

Ms-inb plot for events in the central Asia-western Soviet Union region as recorded by ALPA. NORSAR and LASA. All the earthquakes were from the time interval September-November 9

Wave number domain representation of two interfering signals, one 100X the power of the other, as they would appear recorded by the ALPA array. ' 10a

The same representation as Figure 10a after a wavenuraber filter has removed the effects of the large signal. The second signal is now clearly visible and a good estimate of its power is obtained. job

Spatial coherence exhibited at LASA by a Rayleigh wave from an event in Bai a California. -Q

Spatial coherence exhibited at NORSAR by a Rayleigh wave from an event in Turkey. 12

INTRODUCTION

One of our current tasks in the SAAC in Alexandria

is to evaluate the capability of the network of the

three large LP arrays, NORSAR, LASA and ALPA to detect

and measure the amplitude of Rayleigh waves associated

with P-wave detections reported in the LASA and NORSAR

daily event summaries. There arc two facets to this

work. One is to perform the actual analyses and the

other is to develop und implement suitable processing

techniques so the data may be analyzed automatically

within real-time.

The assumption is made that the SP detection thres-

hold is inherently lower than the LP threshold. Hence,

the SP bulletin is used to predict times at which Rayleigh

waves, associated with each reported event, would be

recorded by the three large arrays» All the LP data from

the three arrays are stored in multiplexed form on a

common tape; the SAAC low rate tape. Each tape contains

approximately 12 hours of data. Software has been

developed to access this tape directly, edit the indi-

vidual channels and perform frequency-domain beamforming

to detect and measure the power of Rayleigh waves. Time

domain beamforming and plotting are available for visual

display.

Preceding page blank

-i-

CONTINUOUS ANALYSIS OF LASA LP DATA

One of the experiments performed was to analyze 50

days of LASA LP recordings using the LASA Daily Event

Summary as a guide. To summarize the results the events

were assigned to one of four azimuth ranges and one of

two distance ranges. These are shown in Figure 1.

The ratios of detected to missed Rayleigh waves

are shown in Figures 2a - 2h, The data have not been

reworked to see whether or not more careful processing

could bring out some of the missed events and so these

results do not necessarily reflect true detection thres-

holds. Also only deep events reported by NOS have been

left out but we must presume that many deep events, not

reported by NOA, are included. Even with these restric-

tions there are some striking comparisons. Detection

ratios for events from the south are poor when compared

with those from the west and northwest. Detection ratios

from the north arc also lower than the west and north-

west. Obviously deep events from South America and Asia

are influencing these results but the difference is

sufficient to suggest that comparisons would be similar

if only shallow events were considered.

This continuous study also provided an estimate of

how long a large event could render the array ineffective.

The energy scattered from huge Rayleigh waves washes

across the array from all directions for a period of

hours and makes the detection of small surface waves

impossible. After a magnitude 6.8 earthquake in Japan

no events could be detected at LASA for a period of

4 hours. Also persistent coherent propagation was detected from an azimuth of around 70°.

-2-

ARRAY COMPARISONS

LASA-NORSAR

As part of the network evaluation, individual pairs

of arrays are being compared by beaming them into an

area which is equidistant from the two arrays in question.

In the case of NORSAR and LASA a very convenient region

is the Kurile Islands area. Both arrays are approxi-

mately 65° away from this seismic area. The structure

of the travel paths is somewhat different. Raylcigh

waves received at LASA travel through the north Pacific

and the Aleutians area, then through western Canada,

altogether a rather complex path. The path from the

Kuriles to NORSAR may be considered a bit simpler as the

bulk of the propagation occurs through the northern

part of the Asian continent. Using Kurile islands events

reported in the LASA Daily Event Summary the LP data

from NORSAR and LASA was beamed and searched to detect

and measure Rayleigh waves from these events. The seismic

events used so far in this study were recorded in the

period May-November of this year. No events were used

previous to this period as the current format for data

transmitted from NORSAR to the SAAC was established

on May 1st.

Figure 3 shows the rab - Ms plots for these data

for both NORSAR and LASA. The events used were the same

in both cases. There are more points visible on the

LASA plot. This is because the NORSAR data transmitted

by the TAL, contains gaps during the analysis period

and so some events analyzed for LASA were not available

for NORSAR.

There is however, a good deal of similarity between

the two distributions, with most of the events occupying

the same relative positions in the two respective plots.

Of the points that are common to the two, most exhibit

a slightly higher Ms value at NORSAR than at LASA. The

difference is typically 0,2 - 0.3 M but some larger

differences can be seen. This may be an effect of the

travel path difference mentioned before.

In Figure 4 the events have been broken down into

magnitude groups with a view to establishing detection/

non-detection ratios. The magnitudes are the body wave

magnitudes reported in the LASA Daily Event Summary.

Rayleigh waves from all ev nts with m. > 5.0 were

detected, as would be expected. In the range m. ■ 4.5 - 4.9,

two events at LASA and three events at NORSAR were not

detected by either array. In each case high amplitude

waves from other events caused such strong interference

that a detection of waves from the Kurile Islands could

not be confidently picked.

The magnitude range mb 4.0 - 4,4 is the range where

we are currently expending the most effort and it can be

seen that both arrays are quite good.

In the suite of events examined both NORSAR and LASA

missed one event each of magnitude greater than 4,0. Both

were of magnitude 4.1 and were different events. We have

no way of deciding whether or not these events had marked

radiation patterns. Also we do not know whether or not the

events classified as missed due to intorference would

in fact have been defected had the interfering signal

not been present. It Must be remembered that our data

source is the I.ASA Daily livent Summary which we must

presume is not complete especially at m. = 4.0-4.1.

Hence even though detect ions were made for all events

in the data set used it cannot be claimed that we

are detecting surface waves from all events with mb - 4•0• "owever we arc encouraged by the results so far.

Although we have not analyzed many events at magni-

tudes lower than mb = 4.0, a few have been included here

to show that surface wave detections from such small

events do occur. More will be added as the work continues.

It would appear that there is no nujor difference

between NORSAR and LASA in the detection of Rayleigh waves

from the Kurile Islands, at least for events with body

wave magnitudes greater than 4.0.

ALPA-NORSAR

ALPA and NORSAR have special interest because they

are closest to the Sino-Soviet region and the ray paths

from a particular event to each array are approximately orthogonal.

Unfortunately there is no highly seismic region

equidistant from ALPA and NORSAR as the Kuriles served

LASA-NORSAR, but if the central Asian area as a whole

is considered then useful comparisons can be made.

Due to the poor quality of the LP data from both

arrays until just recently, only events recorded since

September have been considered. mb - Ms plots for events

commonly recorded at both arrays are shown in Figure 5.

The distributions are similar but there is no indication

of a systematic difference between the two sets.

Many events were detected at one array and not the

other. Over one third fell into this category because

of no data or bad data at one of the two arrays. This situation is improving.

A few are detected at one array only because of

interference at the others. Figure 6 shows an M - m,

plot for all central Asian events recorded by one array or the other, or both,

NORSAR has the capability of detecting small events

in the western and southwestern Soviet Union which

cannot be detected by ALPA or LASA. So far we only have

a small collection and an Ms - mb plot is shown in

Figure 7. Actually one of these points situated at

Ms = 2.95, mb = 4.9 was not well recorded at NORSAR due

to coherent noise coming from the same azimuth. Figure

8 shows the NORSAR beam for this event. The noise is

rather obvious and has been observed on several other occasions.

THE NETWORK

Figure 9 gives a composite M - m, for the central

Asia-west Soviet Union region as seen by ALPA-NORSAR

and a little help from LASA. Some presumed explosions

recorded by I,ASA before August have been added. These

data are continuously being added to and reviewed as

better processing techniques become available. With

refined processing we consistently detect surface waves

associated with events of m, > 4,0 from Central Asia

and the western part of the Soviet Union, Again it must

be cautioned that our data set is small so far and

cannot be considered complete and so we cannot as yet

establish a network threshold for these regions.

SIGNAL PROCESSING DEVELOPMENTS

The design of our processing techniques is influ-

enced by the volume of data to be processed. A fast

frequency domain beamforming approach has been developed

and utilized and operates at approximately 3-4 times

real-time on continuous data. The output is a continuous

listing of coherent detections with velocity, azimuth,

signal power and signal-to-noise ratio. A wavenumber

filtering scheme is included whereby a small signal

can be detected in the presence of a much larger signal.

An example is given in Figures la and 10b. Figure 10a

shows the wavenumber plane for ALPA at a period of

19.0 seconds with two signals imposed. One is from the

southwest and the other, with only 0.01 the power of

the first, is from the northeast. The smaller signal

is not at all visible. In fact time domain beamforming

for the second small signal would give a spurious output

of 8X the actual signal power. In Figure 10b the filter

has been used and the data renormalized. The second

signal is now dominant, the signal power estimate being

90 per cent of the true value. This type of processing

allows us to detect and measure the power of small

Rayleigh waves.

One other study in progress is the comparison of

the spatial coherence of Rayleigh waves recorded at the

three arrays. This study provides a quantitative com-

parison of signal deterioration across each array. Two

examples are shown. The first in Figure 11 is for an

event from Baja California recorded at LASA. The

spatial coherence is estimated in two orthogonal directions,

It can be shown that the loss in coherence in the direction

of propagation is caused by a variation in phase velocity

at the period used and the loss of coherence along the

mean wavefront is caused entirely by azimuthal scattering.

The data can be explained by an angular scatter of + 6°

and a velocity scatter of + 0.1 km/sec. A good example

for NORSAR is a Turkish event and the spatial coherence

estimates are shown in Figure 12. An angular scatter of

1 10° and a velocity scatter of + 0.2 km/sec would explain these data.

Such studies as these have direct application in array

processing as they indicate what type of weighting pro-

cedures should be incorporated to improve signal-to-noise

ratios. There is also an interesting scientific content

as the scattering affects the spectrum of a signal

recorded at a single point and the fact that the phase

velocity cannot be measured uniquely for a particular

period would impose restrictions on structural models

obtained by phase velocity inversion.

-9-

N CENTRALi ASIA

Figure 1. Location of regions used in LASA continuous analysis.

DISTANCE: 61e-100o

DIRECTION FROM LASA: 31S0-450

REGION 1 30

20

10 _J!*VIE,GH 3.S-3.I m. mi WAVES

4.0-4.4 inb

DETECTED

0 —

10

20

30

JMYLEIGH WAVES

MISSED

9

24

16

14

4.5-4.9 inb

10

4 zz^za

S.O-S.S mb > 5.5 mb

zzrz 4

?

- mk RAYLEIGH WAVES INDISCERNIBLE DUE TO INTERFERENCE

(HATCHED AREAS)

Figure 2a. Kayleigh wave detections/non-dctections for region I.

DISTANCE: 25e-60e

DIRECTION FROM LASA: 315e-45e

20 —

REGION 2

in JIAYLEIGN ,MrWAV»

DETECTED

..-

RAVIEIGH WAVES MISSED

3 6 3 ' "fc 4P44«.k J

6 7 / I f

^4.641 iBh 6.0-B.S nib > B.5 Mb

^

1 RAYLEIGH WAVES INDISCERNIBLE DUE TO INTERFERENCE

(HATCHED AREAS)

20

Figure 2b. Rayleigh wave detections/non-detections for region 2.

DISTANCE: 61o-100o

DIRECTION FROM LASA. 450-1350

REGION 3

10 RAYLEIGH

WAVES DETECTED

4.0-4.4 iiib

3.6-3.9 inb

10

20

RAYLEIGH WAVES MISSED

8

j

8

4.5-4.9 mb

7Z1 Ty

5.0-5.5 mh \ >6

2 Z ^-^ s

1

RAYLEIGH WAVES INDISCERNIBLE DUE TO INTERFERENCE (HATCHED AREAS)

Figure 2c. Rayleigh wave dctections/non-detections for region 3.

DISTANCF: 25o-60o

DIRECTION FROM LASA: 45e-135e

REGION 4

20

IQ JtAVLEIGH 3.63J iib WrWAVE8

4.0-4.4 iiib

DETECTED

10

20

30

9

JRAYLEIGH WAVES MISSED

12

y^1

¥^-

4.54.1 iiib 6.0-6.6 Mb

I - I >5B">

2

17

Z

g« J

RAVLEIGH WAVES INDISCERNIBLE DUE TO INTERFERENCE

(HATCHED AREAS)

_

Figure 2d. Rayleigh wave detections/non-detections for region 4.

DISTANCE: 61o-100o

DIRECTION FROM LASA: 135e-2250

20

RAVLEIGH WAVES DETECTED

Ml 3.8-3.9 iiib

REGION 5

4.0-4.4 r^ >».5-4Jmb 5.0-8.5 inb

10

20

30

40

50-

RAVLEI6H WAVES MISSED

8

54

10

31

^

*m

11 10 >5.5 1115

1

RAYLEIGH WAVES INDISCERNIBLE DUE TO INTERFERENCE (HATCHED AREAS)

Figure 2e. Rayleigh wave detections/non-detections for region 5.

DISTANCE: 25o-60o

DIRECTION FROM LASA: 135e-2250

20

10

REGION 8

3.5-3.9 ni|, 4.0-4.4 nib

RAYLEIGH WAVES

DETECTED

10

20

30

40

RAYLEIGH WAVES MISSED

8

35

8

m. 4.5-4.9 fflb

8 5.0-5.5 mb > s.5 mb

RAYLEIGH WAVES INDISCERNIBLE DUE TO INTERFERENCE . (HATCHED AREAS)

Figure li, Rayleigh wave detoctions/non-detections for region 6.

4.0-4.4 Mb

40

50

RAYLEIGH WAVES MISSES DUE TO INTERFERENCE (HATCHED AREAS)

Figure 2g. Rayleigh wave detections/non-dctcctions for region 7.

20

DISTANCE: 25o-60o

DIRECTION FROM LASA: 225°

3.B-3.9 mb 4.0-4.4 inb

-315°

REGION 8

10 RAYLEI6H WAVES

DETECTED 12 12

4.5-4.9 mb

8 5.0-5.B HU

0 4 3 1

RAYLEIGH WAVES MISSED

13

4 j^iJ 1

W, 1

10

20

30 RAYLEIGH WAVES INDI!

(HATCI ICERNIBLE DUi IED AREAS)

: TD INTERFEH ENGE

Figure 2h. Rayleigh wave detections/non-detections for region 8.

< <

CD

(U T) t- O u o

'/I

c QJ

> 01

CJ rH • H ^

O

o •H 3

O

E

U

L0

to

be ■ H

s l/\l

OS < in oi o

4-1 re

o •a ^ o u (U u

** m

£2

$ ^ ^

m

r 5

w

c >

3

o

3

U o

s

CM

u

<S)

m CO CM J-

u 3 M

•H

rW

LASA - KURILES RAYLEIGH WAVE DETECTIONS

4.0-4.4 ni|,

30

20

NO. OF DETECTIONS

10

2 3.5-3.9 iRb

SAAC m|| R/.4GE

34

NO. NOT DETECTED -Mggl^

4.5-4.9 mb 5.0-5.5 »b

6 5.5-5.9 mb

6.0-8.4 mb

r

10 HATCHING DENOTES MISSES DUE TO INTERFERENCE

Figure 4a. Rayleigh wave detections and non-detections for Kurile suite at LASA.

NORSAR - KURILES RAYLEIGH WAVE DETECTIONS

30

20

10

4.0-4.4 mb

NO. OF DETECTIONS

3.5-3.9 mi,

SAAC nib

RANGE

NO. NOT OETECTED

29

fe«!

4.5-4.9 mb

10

?7??%za

5.0-5.4 nib 5.5-5.9 nib

1.0-6.4 mb

y

10 HATCHING DENOTES MISSES DUE TO INTERFERENCE

Figure 4b, Rayleigh wave detections and non-detections for Kurile suite at NORSAR.

NORSAR

x PRESUMED EXPLOSION

ALPA

55

• EARTHQUAKES

4 5 6 SAAC mfi

Figure 5. M - nu. , .._.._..,. ,,._.... „.„..k.,. both ALPA and NORSAR September - November 1971.

- mu plot for central Asian events recorded at

< 5

SS

o -d u o o ü ^ in 4-1

C • 0) t-H > h- 4) Ol

rH C rt (-,

■H O '/) J3 < E

o « O

4-t C i o 'J >-.

Hl ^ x> O E ^ o •M ß. 0 OJ

—I LO

E LO cd

i C

o ■ ^

C o ca

3 <

c

5

> o

c u u

o S x; 4-1 3 O 1/1

T) C ra

C v- 0) 4-1 I VI Oi 4» < St^ u. ^ o o z:

4-1

4-1

O T)

a-3

ja o E u

■ >-

1/1 VI S 4-1

c • o

l-~ >

Ü

3

•H C U. 3

C o

s w

f^JMlMM». •

2

i

e o •

<* c V

M > c u

• H

ca rt bC-H n I« P. V) O 3 up: & 0) V) « o

•H 15 o c rt ^ e c o V u

Ä 0) o > u ca

to CC

•H M »•H O Q

EOS rt 0) n

x> oe

a: c <.H

O tfl ^ 6

(ü n

3 M 0)

•H JS

5

i u

• o I <Xl

•H < 01

' -a a

>H f—I

C < > O LT, S- u a: o

o t-1

o ^ e

W >. Ü

O 4-< > "3 O O E

-a o u u u o o *-*

o o 4-> l-i ^■ O Ü

rH C S ^•2 t/,

!«■>

o ,s «^ W 4) .

M

COO Ü I- JS J3 I- O t-l E 3 W O

M O —• O X 2S < 12

s HI/

3.6 Km/sec

PEAK POWER = I.SxIO9

CONTOURS db DOWN FROM MAXIMUM

I'Lgurc 10a. Wave number domain representation of two interfering signals, one 100X the power of the other, as they would appear recorded by the ALPA array.

12

12

/3.6 Km/sec

PEAK POWER = 1.7*107

CONTOURS db DOWN FROM MAXIMUM

Figure 10b. The same representation as Figure 10a after a wavenumbcr filter has removed the effects of the large signal. The second signal is now clearly visible and a good estimate of its power ' obtained.

i s

Cft

to UJ

XI 60

13 Ü

* C a» u

o o CM u -H C M ü n i- u

o en

« •H C

a) 0.+J

•-0 c 0)

• > i—I OJ

c

3 E M O

■H I- U, '-M

/ 33N3U3H03

1.0

^010 oa>ooo 0 « 0

O o 00

CM

** 0.6 CJ

0.4

A A

O NORMAL TO WAVEFRONT

A PARALLEL TO WAVEFRONT

A A A

-

20 40 60 80 100

SEISMOMETER SEPARATION (Km)

l;igurc 12. Sp.itiiil coherence exhibited at NORSAR bv a Rayieifth wave from an event in Turkey.