'■ra^BWTOwww»:""*«! yw. 1111 jpit|ijpa^wBfi5}(|!iBRii!. i p

# J

o Oi CO

'S?'!; '

T

SEISMIC AMPLITUDE

AND WAVEFORM RESEARCH

R. S. HART D. M. HADLEY G. R. MELLMAN R. BUTLER

'■ '"-<

SGI-R-79-012

S\> ^,^6 ^Mr

J'.

FINAL TECHNICAL REPORT

SPONSORED BY

DEFENSE ADVANCED RESEARCH PROJECTS AGENCY (DOD)

ARPA ORDER NO. 2551

■ \

f i o

i

■•.ftctsT

This research was supported by the Advanced Research Projects Agency of the Department of Defense and was monitored by AFTAC/VSC, Patrick AFB, FL 32925, under Contract No. F08606-79-C-0009.

The views and conclusions contained in this document 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 United States Government.

APPROVED FOR PUBLIC RELEASE, DISTRIBUTION UNLIMITED

JM October 1979

SIERRA GEOPHYSICS, INC.

150N. SANTA ANITA AVENUE . ARCADIA, CALIFORNIA 91006 . (213)574-7052

80 4 28 120 ■ ■ ■■ ' ;.■.■ -V 1 ' J ..-.^^^B "

■ -TW^WWWIW^»?»^ » P5B»T.-m^^»WWWMW«U!«p;«^^^^^^^^

AFTAC Project Authorization: VT/9710

ABPA Order: 2551

Effective Date of Contract: November 22, 1978

Contract Expiration Date: September 30, 1979

Contract No: F08606~79-0009

Principal Investigators and Phone No:

Dr. Robert S. Hart

Dr. David M. Hadley

Dr. Rhett Butler

(213) 574-7052

Program Manager and Phone No:

Mr. Michael J. Shore

(202) 325-7581

i

^J;

'

}

■ mmmmmmmm immmmmm^m^mmm^f^ii

SGI-R-7y-0I2

i

SEISMIC AMPLITUDE

AND WAVEFORM RESEARCH

R. S. HART D. M. HADLEY 6. R. MELLMAN R. BUTLER

■ ...

FINAL TECHNICAL REPORT

SPONSORED BY

DEFENSE ADVANCED RESEARCH PROJECTS AGENCY (DOD)

ARPA ORDER NO. 2551

This research was supported by the Advanced Research Projects Agency of the Department of Defense and was monitored by AFTAC/VSC, Patrick AFB, FL 32925, under Contract No, F08606-79-C-0009.

The views and conclusions contained in this document are those of the authors and should not be interpreted as necessarily representing the official policies, either expressed or implied, of the Advanced Rasearch Projects Agency, the Air Force Technical Applications Center, or the United States Government.

APPROVED FOR PUBLIC RELEASE, DISTRIBUTION UNLIMITED

October 19 79

BfilWt mmmm

PipWBps^PWPw^BJi mmmmmm»mJmi9 SifHPfPSPPiifSPW^PiP^^

i

\

UNCLASSIFIED SECURITY CLASSIFICATION OF THIS PAGE (Whn D«'» Enf.f.d;

REPORT DOCUMENTATION PAGE t. REPORT NUMBER

2 GOVT ACCESSION NO

ßjtAälz 1 4. TITLE fand Sublll/eJ

Seismic Amplicude and Waveform Research^

READ INSTRUCTIONS BEFORE COMPLETING FORM

3. RECIPIENT'S CATALOG NUMBER

£2 6. Ty,PE OF REPORT 4 PERIOD COVERED

Final Report/^

&\ 7. AUTHORCJ'

R. S. Hart, D. M. iladley, G. R. Mellman, R. Butler

T FiflPORMlNG ORGANIZATION NAME k«" -JCRESS

Sierra Geophysics, Inc. / ,< ^ > Jj, 150 N. Santa Anita Ave. h>J _L^ J

Arcadia, CA. 91006 ___^__—

6. jER^RMWO^ROnREPORT NUMfMLR

SGI-R-79-£/l2 / ^B—CONT-RACT O'R GRAN- fcUMBER^)

."tg] FOSöyoe-yg-^sZ"- 5^S

—4—1—_ ' f_ ,_-.^ L. ";-;:-.. ,; „ ,,,, ■ t ,- • T i >; x .10.--PROGR-AM-EL-EMENJ..RROJ£|CT,_T.ASÜ AREA 4 WORK UNIT NUMBERS

II. CONTROLLING OFFICE NAME AND ADDRESS

VELA Seismological Center ill Montgomery Street Alexandria. Virginia 22314

ARPA Order No. 2551

12. REPORT DATE October 25, 1979

14. MONITORING AGENCY N AME & AüÖRESSm dliltren, Iron, Comro/Hn« Ollice)

(Uj'iS Ort 11 ^

13. NUMBER OF PAGES

15. SECURITY CLASS, (ol this report)

6 7£ t'' l^- 15« DECLASSIFICATION DOWNGRADI N G

SCHEDULE

Vs. DISTRIBUTION STATEMENT (ol this Keporl)

Approved for public release, distribution unlimited,

17. DISTRIB UT.ON STATEMENT (ol t*. .t.„ac, enter.* In BlocH 20. il älllerent Irom Report) , / , , I j

10) ULjJUlhn <,-., n.UnjEvfk^

18. SUPPLEMENTARY NOTES

19. KEY WORDS (Continue on reverie side il necessary and identity by btocH number)

Underground Nuclear Explosions Seismic Source Functions Attenuation Body Waves . .LV Frequency Dependent Q niX (CUJtH

20. ABSTRACT (Continue on rave,., side il necessary and Identily ^ Mock number,

^This technical report summarizes the/research done in a 12 month program directed toward understanding observed^ and yield biases at the Nevada Test Site This program has necessarily incltfded work on seismic source functions, variations in attenuation, frequency dependence of Q, regional amplitude and waveform variations, and shallow structural influences 01 seismic wave propagation. The report is divided into six sections, each describing an individual research effort on one or more of these issues. —-;, -f# 4>. -«-

pQ FORM 1^2 EDITION OF 1 NOV 65 IS OBSOLETE UNCLASSIFIED O /Ö 0 U>J' S^- SECURITY CLASSIFICATION OF THIS PAGE (Whin Data Entered)

•»■•)R.'aiMBBBneii*»Wj. ■.

'*f''*SiSliiS^^

TABLE OF CONTENTS

Introduction

/l.) Seismic Source Functions and Attenuation from Observations of Pahute Mesa Events I N . . . .

t-^y^t) Observational Studies of the Piledriver and ■ -J Hard Hat Tests *, v

vf^)1 Frequency Dependence of Seismic Attenuation: A Mechanism to Unite Amplitude and Spectral Ratio Techniques I ■)

IV.v..i Estimation of Receiver Functions for Stations Located at the NTS i

■

4.1 Theory . . .

4.2 Application to Data

4.3 Interpretation of the YF Receiver Functions . . lv\

V. .^Studies of Outgoing Seismic Energ> from Yucca Flats, NTS ' to*,£.\

V^H VI. -.^Amplitude Studies of the Stations 0B2, RKON and HNME Relative to the WWSSN Mean «,

References

- i-

Page

1

24

24

30

48

54

■ »

.^^;^;w^«.r,^rf^^^

::

INTRODUCTION

U

During the past ten months, Sierra Geophysics has been conducting a

research program aimed at understanding the IIL biases observed for under-

ground tests located at the Nevada Test Site (NTS). This report summarizes

this ARPA-supported research which was monitored by AFTAC/VSC.

The research program was divided into three principal tasks. The

first task was an analysis of both incoming and outgoing seismic energy

at NTS to determine how the waveforms are affected by local geologic

structure and intrinsic attenuation. The second task involved a theoretical

analysis of wave propagation with models representing the complex, three-

dimensional structure at the NTS. The final task was to combine the results

of the first two tasks and to estimate the sensitivity of the data to vari-

ations in the structural model for the region or parts of the region.

The program summarized in this report does not represent a final answer

to the question of NTS m, bias. Additional work on this problem, including

the acquistion and integration of seismic data from other Special Data

Collection Systems (SDCS) stations (NT-NV, GQ-NV, etc.) and the inclusion

of additional structural information (expected shortly from studies by Herrin),

is needed.

This report is divided into six sections. The first two sections

summarize research on teleseismic source functions appropriate for

NTS events and on teleseismic attenuation measurements. These first two

summaries have been extracted from the Quarterly Reports SGI-R-79-002 and

SGI-R-79-003.. The third section deals with on-going studies of the

frequency dependence of seismic attenuation and the implications for yield

■"^•mviifierw

mmm/ti¥'"^- ■ ^mvimmm'm^pm*

determination and discrimination. The fourth section discusses research

conducted on both incoming seismic energy recorded by the SDCS stations

at Yucca Flats and at Climax Stock and outgoing energy from tests at

Yucca Flats. The fifth section is a discussior of model studies of

wave propagation at Yucca Flats and a comparison of those results with

the. conclusions reached in the data analysis of the YF-NT data (Section IV)

Finally, the last section reports on amplitude comparisons between the

SDCS stations 0B2-NV, RK-ON, and HN-ME and WWSSN stations within the

continental United States.

i

I

/

i,4- •

PMPpiiiliWi!! -•>?pTOy.Wmiwj^p||ip||||^

distance the initial seismic energy represents diving rays and hence is

suitable for comparison with more distant regional and teleseismic

observations. Using a modified Von Seggern and Blandford (1972) source

representation, and including near-field terms, it has been possible to

obtain source functions which not only accurately model close-in records

but also match teleseismic observations. Having once defined the explosion

source description, it is a straightforward task to determine the effective

t* for teleseismic observations without the usual ambiguity of what are the

source influences as opposed to the anelastic effects. For WWSSN short

period observations of these events, we obtain an average t* of about 1.3

for compressional waves with a scatter of about ±0.2. There are systematic

azimuthal trends in the observed t* values which are not strongly correlated

with the Silent Canyon caldera but may be correlated with part of the central

Rocky Mountains. It is not possible at this time to rule out systematic

receiver function biases as the cause of the amplitude variations. A

principal, although for present purposes not critical, limitation on

the source function determination made in this study is the uncertainties

in the precise crustal structure and seismological properties along the 8 km

paths between the events and the strong motion sites. Since these

uncertainties directly affect the resolved source function, this structure

needs to be more precisely defined in the efforts to reduce the observed

yield variations at NTS. A complete description of this research project

is contained in the Sierra Geophysics Quarterly Technical Report "Seismic

Source Functions and Attenuation from Local and Teleseismic Observations of

the NTS Events Jorum and Handley" (SGI-R-79-002).

U: •!- KJ,..W,

«mmm

■■■ ■ ' ■ ■ . ■■

II. OBSERVATIONAL STUDIES OF THE PILEDRIVER AND HARD HAT TESTS

This study represents the second part of a project to define both the

teleseismic path-averaged attenuation operator appropriate for sources

located at NTS and the deviation from the average associated with distinct

geologic structures. In section I of this report we summarized the first

part of this program. In this section we summarize research on the Pile-

driver and Hard Hat tests conducted in granite at Climax Stock. A detailed

description of this work may be found in the Sierra Geophysics Quarterly

Technical Report "Seismic Studies of the Nevada Test Site,," Section 3

(SGI-R-79-003).

In this study, we have tried to relate the near-field and teleseismic

data for the test Piledriver (yield -56 kt; Springer and Kinnaman, 1971).

Data from a second event. Hard Hat (yield -5.9 kt) were also examined as

a check on the Piledriver results. Hard Hat was located within the Climax

Stock and was separated in distance from Piledriver by about three hundred

meters. Although this shot was much too small to be recorded by the WWSSN

array, both Hard Hat and Piledriver were well recorded by the short period

Benioff instruments operated in Southern California (A - 500 km) by Caltech.

This research project combined and interrelated near-field data from Pile-

driver, scaling relations for a 56 kt source, teleseismic waveforms from

Piledriver, regional amplitude data from both Hard Hat and Piledriver and

RDP's from Hard Hat. Unfortunately, these data sets are not internally

consistent and the results must depend, to some extent, on which data are

heavily weighted in the analysis.

I

If the velocity records recorded at the distances of 204 m and 470 m

from Piledriver are employed to calculate an effective RDP for the source,

we find that the teleseismic short period P waves predict a t* of 1.3.

However, some question remains as to the degree in which these near-field

records have been contaminated by non-linear effects. The calculation of

an effective RDP implicitly assumes that such contamination is negligible.

This source description was also compared via several source scaling

techniques to the experience from other tests. On the basis of this

comparison, the RDP strength derived from the close-in velocity records

appears to be a factor of 2 too high. This factor is roughly equivalent

to that obtained from a comparison of regional amplitude data recorded in

southern California from the Hard Hat and Piledriver tests. The reduction

in the RDP strength would reduce the predicted t* from 1.3 to about 1.0.

The present information does not provide a conclusive basis to decide which

data set should be given the most weight. For periods of about 1 sec,

the present data only constrain t* to the range [1.0-1.3] for sources

located at Climax Stock.

: ■ ::■ ;-;■■.vv-'v ■ ■ ■■■■ ■.■■ :■■■ ■

L

III. FREQUENCY DEPENDENCE OF SEISMIC ATTENUATION: A MECHANISM TO UNITE

AMPLITUDE AND SPECTRAL RATIO TECHNIQUES

The attenuation of short-period body waves, commonly described by

the parameter t* (the ratio of travel time to Q) , is an extremely important

factor in both the detection problem and in the source description analysis

of underground nuclear explosions. The attenuation of short period seismic

energy has been the source of major controversy within the seismological

community. A large part of the disagreements may be traced to differences

in technique and the frequency bands being considered. The time domain

determinations of t* are dominated by .5 to 2 hertz energy, the frequency

domain estimates have concentrated on higher frequencies and have resulted

in smaller t* values.

We have approached the problem of defining a useful attenuation model

by first modeling the time domain waveforms, and then narrow-band filtering

(center frequency = 1, 2, A, Hz) the observations and the synthetics, and

by comparing spectral amplitudes. This dual approach provides both a time

domain and frequency domain check on the interpretations. This study was

initiated primarily to investigate the trade-offs between the frequency

behavior of the attenuation operator and absolute (M. ) amplitudes of

synthetic teleseismic waveforms. The study to date has been necessarily

limited to just five tests recorded at 0B2, HNME and RKON. Within this

report we will restrict the discussion to the simplest event and to just

the RKON and HNME data. The generation of the synthetic seismograms

requires a source description and a t* operator. Following previous work,

we have used a von Seggern and BJand ford (1972) source parameterization

.i i *■ in«

., .... •;,-...

u

.

8

and have adjusted the rise tin*, on the basis of the Piledriver/Hard Hat/

Shoal experience, to be approxi.ately representative of a 100 Kt underground

explosion (K-7.6).

Figures 1 and 2 show the raw seismograms (top traces) for a Russian

underground explosion. This event was located at the Northern Novaya

Zendya test site and the date was Septe.ber 1, 1977. Traces 2-4 are narrow

band-passed (.ero phase shifted) filtered records confuted after the raw

data were corrected for the Instrument response. The log spectral amplitude

Plots, Figure. 3 and 4, show the instrument corrected data. For HNME very

little 4 Hz energy arrived at the P-wave onset. However, for RKON the 4 Hz

energy arrived with the 1 sec P-wave. Next, synthetic seismograms were

computed for these two records. The t* chosen for each station was based

on the observed amplitude variations for Russian underground explosions,

i.e. RK0N/0B2 = 4 and HNME/0B2 = 1.7 (see Section VI). Assuming t* =13 0B2 ■L,J>

these observations approximately translate to t^ . o.65 and t* HNME . 1.05,

The PP reflection time was adjusted in order to obtain the best visual fit.

Finally, the synthetics were superimposed on the noise preceeding the

observed records and the resulting seismograms analyzed with the same

program used on the actual data, Figures 5 a,d 7 P u , rigures i a.id 7. Perhaps as expected,

these synthetics visuallv f-ii- n^ u visually fit the observations quite well. Note also

that the chosen t* valueq (anrt ha„~„ \ values (and hanca n^) predict the observed amplitude

ratios between RKON and HNME (data ~ ■* / ... UMHE »data - 3.4, synthetics . 3.7). For HNME

the amplitude observed in farh *-„„ in each frequency band is fairly consistent »Ith

tbe synthetic. The spectral plot. Pi8ure 5. indicates that the synthetic

as sli8htly depleted at freouencies around 3 HZ. However, the synthetics

data. The spectral pint, Flsurc 3, shows that the discrepancy is so^hat

. . — -..—

CODE: 770109HN-ME yMqx= 100.812 RATIO

■

i- CENTER FREQUENCY: I HZ. XMnX= 79.453

.79

CENTER FREQUENCT: 2 HZ. XMflX= 12.941

O

.13

O CENTER FREQUENCT: 4 HZ. XMflX= 0.521

O

ImmilmL ß .002-.005

Figure 1: Waveform recorded at HNME from a Russian test. The lower three traces have been narrow band-pass filtered after correcting the top trace for the instrument response. The ratio is the maximum amplitude of the filtered trace divided by the maximum of the recorded record.

. . • -,.. .^ h-jM. mim^tmit*

■ 'i'""-1 ''■'"' ■■ ' •■ \ ... .-.:a m

10

o

10G SPECTRUM PLOT

1.000

770109HN-ME

a 0.100

CL

a

a

U LU 0.

0.010

0. 001

5. Q

0 FREQUENCY. HZ

O

o

::

SL0PE=10. «« (-0.6081/HZ

Figure 2: Fourier transform of the trace shown at the top of i-igure 1. The spectrum has been corrected for the instrument response. The dashed line is a least squares fit to the spectrum over the range [.5-5 Hz].

.......>,, '77",r'r:~TT ——■ -r ■:—— r-r-

I 0

1 "' ._

11

TODE: 770l09RK-nN XMRX- 342.566 RATIO

^i I \n/\/AAj\I\/\fA^^

CENTER FREQUENCY: 1 HZ. XMRX= 127.396

,37

CENTER FREQUENCY: 2 HZ. XMRX= 55.931

O

'vV\/VAAA^^v\Ay\/WVA^/yV .16

CENTER FREQUENCY: 4 HZ. XMRX= 3.41Ü

( i

^^^^■«^-■■■^ **I^**>^^/^^^^^*~~^\^*I\S^/J |i^^ • 01

Figure ">,: Waveform recorded at RKON from a Russian underground explosion.

I ■*——-—- -• ■*—- ■

•—— ' ■-■

(J

LOG SPECTRUM PLOT

1.000

a 0. 100 3

s: cr

cr cr •— u UJ a. ^ 0. 010

0, DO I

770109RK-C1N

3.0

FREQUENCT. HZ

12

4.-Ü 5. 0

5L0PE=10. mm (-0. 500)/HZ

Figure 4; Fourier transform of the trace shown at the top of Figure 3.

ij

I 3

■ ■ .J. ,.'■; ^'■, s I.—

13

U HNME-SYNTHETIC XMflX= 89.027

RATIO

CENTER FREQUENCY: 1 HZ. XMflX= 73.732

• 32

U

o

CENTER FREQUENCY: 2 HZ. XMflX= 11.245

,125

CENTER FREQUENCY: 4 HZ. XMRX= 0.172

ifiilN^^ .002

Figure 5: Synthetic waveform for the'record at HNME. The waveform has been added to the noise preceding the actual P-wave arrival at HNME. The synthetics were calculated with a von Seggern and Blandford source (K=7.6, B=1.5), t* (independent of frequency) = 1.05 and pP was lagged .25 sec.

"■M!.... ■ ■■■^ '■ ■'■■^ 3BSBSB»!^WI*» ^

f.. —. • .- .—-. mum, i v .■..^M , ■, .-d-i .„„„.„^ .-... .il« i i Vi il -^-1 i. tlli'tfiMr ; ■ ',ai

14

1.000

o

u

Q 0.100 3

(L.

er

er cc t—

u UJ Q.

0.010

0.001 l 0. 0

FREQUENCT. HZ

Figure 6: Fourier transform of the top trace shown on Figure 5.

1

J

0

' ' i'—nwa

■'.:, —^^_:i" 1 .. -■uj—^MM—MMMü oM

CENTER FREQUENCY: I HZ. XMflX= 194.178

CENTER FREQUENCY: 2 HZ'. XMflX= 52.890

CENTER FREQUENCY: 4 HZ. XMflX= 0.625

15

RKON-SYNTHETIC XMflX= 329. 170 RATIO

_J1

.59

i |

.16

/(/yywwM/i^^ \vwwvVVWW(/Wi/yVl|lHV^^ •002

y Figure 7; Synthetic waveform for the record at RKON. The source model

Is identical to that described on Figure 5 t*=.65.

mimLüMm^'r "^m ' ' y Rif..pt.i.'i;'.i-'.'p..-'iji

B u 16

U

1.0ÜO

a 0. 100 ID

a. cr

a or h- (_»

a. in 0.010

0. GDI 0.0 1. 0 2.0 .3.0

FREGUENCT, HZ

4. C 5. 0

Figure 8; Fourier transform of the top trace shown on Figure 7.

.s&i^mwm

17

(J

4 Hz energy is routinely observed (Der and McElfresh, 1975). However,

decreasing t* to accommodate the 4 Hz data would strongly violat' the

observations at 1 sec. The RDP's at 1 sec are reasonably accurate and the

teleseismic amplitudes require an operator with the characteristics

described by t* ~ 1. One reasonable compromise discussed by Liu,

Anderson, and Kanamori (1976); Kanamori and Anderson (1977); Anderson,

Kanamori, Hart, and Liu (1976); Anderson and Hart (1977, 1978); and

Hart (1977) is to move the high frequency edge of the absorption band

into the range 2.5 Hz. ».. have coded the Minster (1978a, b) description

of an attenuation operator defined by an absorption band model. The

relative spectral amplitudes of this operator as a function of T (~l/2Trf), m

are shown in Figure 9. The attenuation for 5 possible postions of the

absorption band and t* =1.0 are shown. Note that this range affects the

1 Hz data by about a factor of two but can potentially alter the 4 Hz data

by many decades. For the data presented for RKON, moving the edge of the

absorption band to about 2 Hz (T~0.08) would more than account for the

high frequency energy. As an example, synthetics for RKON (t* = .75)

and for HNME (t* = 1.15) for T =0.08 are shown in Figures 10 and 11. m 0

A comparison between either of the band-passed records or the spectral

plots. Figures 12 and 13, and the actual data show that the synthetics

now have too much high frequency energy. For this single example it

seems that the position of the absorption band is not constant: HNME

is fairly well modeled with T ~ .001 for RKON T < 0.08. mm

Currently there are two distinct methodologies in the analysis

of seismic attenuation: Amplitude data from both short and long period

( \

'

HI ii

17

4 Hz energy Is routinely observed (Der and McElfresh, 1975). However,

decreasing t* to accommodate the 4 Hz data would strongly violate the

observations at 1 sec. The RDP's at 1 sec are reasonably accurate and the

teleseismic amplitudes require an operator with the characteristics

described by t* ~ 1. One reasonable compromise discussed by Liu,

Anderson, and Kanamori (1976); Kanamori and Anderson (1977); Anderson,

Kanamori, Hart, and Liu (1976); Anderson and Hart (1977, 1978); and

Hart (1977) is to move the high frequency edge of the absorption band

into the range 2.5 Hz. We have coded the Minster (1978a, b) description

of an attenuation operator defined by an absorption band model. The

relative spectral amplitudes of this operator as a function of T (~l/27rf) m '

are shown in Figure 9. The attenuation for 5 possible postions of the

absorption band and t* = 1.0 are shown. Note that this range affects the

1 Hz data by about a factor of two but can potentially alter the 4 Hz data

by many decades. For the data presented for RKON, moving the edge of the

absorption band to about 2 Hz (T~0.08) would more than account for the

high frequency energy. As an example, synthetics for RKON (t* = .75)

and for HNME (c* = 1.15) for Tm = 0.08 are shown in Figures 10 and 11.

A comparison between either of the band-passed records or the spectral

plots, Figures 12 and 13, and the actual data show that the synthetics

now have too much high frequency energy. For this single example it

ssems that the position of the absorption band is not constant: HNME

is fairly well modeled with T ~ .001 for RKON T < 0.08. m m

Currently there are two distinct methodologies in the analysis

of seismic attenuation: Amplitude data from both short and long period

«»»lull»—! ii iwy

18

Rttenuation vs Tau m, t =1

1.000

UJ a 0.100

Q-

en

cr cr t- u a. to

0.G10

0. 001 0.0 1.0 2.0 3.0 4. Q 5.C

FREQUENCT. HZ

Figure 9: Spectral amplitudes of attenuation operators described by an absorption band model. Five possible positions of the high frequency corner of the band are shown.

tifalk-,^!.^.....:.!'. •— ' '— -. : 'rrnr

-"^— ——""»""-^"^

19

RKON SYNTHETIC XMRX= 9GS.010 RATIO

CENTER FREQUENCY: 1 HZ. XMflX= 351.182

.36

CENTER FREQUENCY: 2 HZ. XMRX= 164.593

——^AA| t / wX/v*'^^—-^-^—- .17

CENTER FREQUENCY: 4 HZ. XMflX= 22.319

/ ,~VSÄA^ V ■;'' i / ■ 11 ll l/VWVVVww- ,02

Figure 10: Synthetic wavefotms for frhe record at RKON. operator used is described by t*-.75 and :'■ model id the same as described on Figure S™

The attenuation 08. The source

—"-V-- .■'--. .-'. ..^I" ■:. . '

■ r^^ '■'.■. „^

20

•^

000

o

a 0. 100

a:

cr or t— <_» LU Q.

0.010

0.001

2.0 3. 0 4. 0 5.0

FREQUENCY. HZ

li

Figure 11: Fourier transform of the top trace shown on Figure 10.

«-'-""."i -■'

u

a

o

;:

•»^ 11" '31

21

HNMU-SYNTHETIC XMnX= 316.080 RATIO

CENTER FREQUENCY: I HZ. XMflX= 154.527

.4'8

CENTER FREQUENCY : 2 HZ. XMflX 124

—-Wl!1 ,17

CENTER FREQUENCY: 4 HZ. XMflX= 4. 814

n Figure 1.2:

^^||P ^\VV\AA«. -*^WV^AA»* w^^vw^mi .015

Synthetic waveforms for the record at HNME. The attenuation operator used is described by t*=1.15 and - =.08. The re- flection time for pP was changed to 0.35 sec. The source model is the same as described on Figure s.

^m

• -11 -:■■-- -■■>■ ■ ■-'■■- IIHKHM

22

L

1.000

a 0. 100

a. cr

en oc i- u UJ

10 0.010

0.001 5. C

O FREQUENCY. HZ

O

Figure 13: Fourier transform of the top trace shown on Figure 12.

■ -■■'W**:- -"■-v-,'vv

23

seismographs (which are closely related to HL ) and spectral ratios. The

amplitude data for each source region and each site have been shown by

Butler (1979) to be quite stable. For the first-order estimate of nL

this amplitude data is crucial. However, towards the goal of modeling a

range of seismic bands (i.e., 1, 2, and 4 Hz) the spectral slope data

should be incorporated into the absolute level estimates at ~ 1 Hz. The

above paragraphs discuss a smooth model that can connect the 1 and 4 Hz

observations. Since the absorption band model represents a temperature

activated process, we expect some relationship between temperature, the

position of the band, and t*, i.e., high temperatures imply large

attenuation and short diffusion times (small T ). As discussed above, m

this model is well founded in previous theoretical works.

The limited modeling study presented above should not be interpreted

as demonstrating the existence of frequency-dependent attenuation. The

principal objective of this study has been to examine for a few simple cases

the proposition that a frequency dependent Q operator can unite the diverse

views of t*. Consideration of Figure 9 and the results from the modeling of

a simple source strongly suggest that a frequency-dependent attenuation

model can unify the current divergent views.

©

24

IV. ESTIMATION OF RECEIVER FUNCTIONS FOR STATIONS LOCATED AT THE NTS

4.1 THEORY

The method that has been used in estimating receiver functions is a

two stage procedure. The first stage uses deconvolution and complex log

spectral stacking to produce transfer functions that map seismograms re-

corded at a reference station into those recorded at other stations. The

second stage uses a Minimum Entropy Deconvolution method (Wiggins, 1978)

to estimate receiver functions for all stations, including the reference

station, from the transfer functions obtained in stage one.

In order to obtain a transfer function relating near receiver re-

sponses of a reference station to another station we must remove the

effects of the source and near source structure. This may be done by

deconvolution If the apparent source, including near source structure

effects, is the same for both stations. To ensure this condition is met,

we must either use stations that are close together, and thus have

similar source receiver azimuths and ray parameters, or use non-directional

sources. Further, due to the presence of near receiver lateral inhomogene-

ities, we should limit events in any stack to a relatively narrow azimuth

range.

Assuming that the above conditions are met, deconvolution of the

reference seismogram from the seismograms recorded at other stations will

effectively remove source, instrument and some attenuation effects. Let

~i —i So (w) and s (to) be, respectively, the Fourier transforms of the seis-

mogram of the ith event at the reference and jth station. The deconvolution

for the ith event and jth station is just cT"!" (co) = s";, (u)/^1 (w)

—▼ T-—~ -■■--' >-■

. -^i-»-.:^ .. ,,..-~-^^.:--.-_-._;J—

25

We may now produce an average transfer function tj (w) for the jth

station relative to the reference station by a complex log spectral stack.

log t. (Cü) 1 n - • -L log d J (u)) i=l J

(2)

(■)

It should be noted that the same formal estimate for t (w) may be

obtained by stacking the log spectra of the seismograms first and then

deconvolving the stacked seismograms. However, by performing the deconvo-

lution first it is possible to avoid the phase unwinding problems that

usually plague complex log spectral methods. While there is no problem deter-

mining log | y(w)| = ± £ log | cTJ (w)|, determining the phase is somewhat

problematic, due to 27Ti ambiguities in the complex log function. However,

by deconvolving first and detrending the dj (co) over a specified frequency

band, which corresponds merely to an absolute time shift of s!(t) with re-

spect to so(t) it is possible to concentrate the d^ (W)in one half of the

complex plane or less for a given a). Phases may then be averaged and pro-

blems of — phase shifts are avoided, since the same 2m absolute phase

shifts are ignored for all n deconvolved signals.

An alternative approach to producing the functions "t.(w) is the stand-

ard deconvolution procedure. This involves determining t1 (a)) that minimizes

the fünctionals e,, = Z I I sT1 (a)) - t" (uti sT1 II ^ TU«-^ J i ' ' 8 j ^ ^ ^ s 0 II • There are several advantages

to using the log spectral stack, as opposed to standard deconvolution methods.

The first is the relative insensitivity of the log spectral method

to absolute timing information. This permits higher frequency information

to be retained than is possible with conventional methods. Also, while

■—r—■!>. '. i ..-.■- . • t ju nawwimi m»

26

u..;

standard methods are effective in dealing with random additive noise, the

principal noise in many seismograms is not stationary additive noise but

rather signal controlled noise, Such noise may be modeled by convolution

in the time domain. Since time domain convolutions become additions in

the log spectral domain, we see that log spectral stacking tends to effec-

tively cancel such noise processes. Causes of such noise include amplitude

and moveout changes of later arrivals due to local near receiver inhomogene-

ities, azimuthal changes from event to event, and changes in the apparent

source from station to station for a single event.

The transfer functions t (tu) obtained from the deconvolution and

stacking procedure contain not only the desired receiver functions,

^(w), but also the inverse of the reference station receiver function,

r (w) thus, we may write

t (w) = r (co)/r (w) j > 1 J J *■'

to(a,) = 1

(3)

O

In order to recover F (u)), and hence 7. (OJ) we must impose an additional con-

straint on the problem. The constraint we shall use is that the r.(t) be,

on the average, as simple or "delta-like" as possible. This may be done

by minimizing a weighted varimax norm

n V = E w'.v = j w

i=o i

f< (t)dt I ,■

V-i (tOdt)' (4)

G

', '

■ -

— ■

27

This procedure, known as Minimum Entropy Deconvolution (MED) and the

properties of the varimax norm have been discussed by Wiggins (1978). In

addition to producing the "simplest" possible estimates of r ...r , the

MED procedure has the property of rejecting incoherent high frequency

energy. Thus, while the ratios of T.(w)/r (w) are fixed by the T-'s, the

absolute amount of energy in any frequency band for r ...r is determined

by the coherence of the T.'s in that band. In this way, the KED process

aids in stabilizing the results of the deconvolution and stacking procedure.

It may be shown that minimization of the functional in equation 4 is

equivalent to solving the system of non-linear simultaneous equations

O

D

V w w „

i i u. (5)

where Ui =fT± (t)dt, * denotes convolution and * denotes correlation.

These equations may be solved for r iteratively, either in the frequency

or time domain. Wiggins had adopted a modified Levinson recursion algorithm

in order to obtain a time domain solution. However, since we expect that

the time duration of r (t) may well be comparable to the duration of the o

r.(t), we have adopted a frequency domain solution. Through the use of FFT

algorithms, it is possible to save considerable computer time in this manner.

In the frequency domain, we may express equation 5 as

- * (uO

W — A I i Pv i 2 i

= Ui

(w) ti((jiO

o I Vi v

v. 1

((0)^(0))

Q

-■■ i ■■ ■■-! ,; . ..

28

O

3 * where R^t) = ri (t) and f denotes the complex conjugate of f. This

equation may be solved iteratively by first assuming r (t) = 6(t-t ), and

hence R^t) = t.3(t-to). We may then solve for r^) and in this way

generate a new estimate of R.(t). In general, this procedure must be

repeated only 4 or 5 times before convergence is obtained. By replacing

Vt) with R.(t) H(t), Ii(t) is a heavy side function, at each iterative step,

we may introduce some semblance of causality into our solution, without,

in general, adversely affecting the convergence rate.

It should be noted that, while the minimum obtained in iteratively

solving (7) is not unique, a useful minimum may easily be obtained by

adjusting the weighting functions w^ m general, however, experience has

shown that the solution obtained is not critically dependent on the choice

of w^s. It should also be noted that, except for possible problems

with multiple minima, the entire stacking and tfED process is invarient to the

choic. of which station is to be used as the reference station. That is. with

proper choice of w , the same set of r. functions is obtained, independent

of the choice of the reference station.

It should also be noted that there is no guarantee that the simplicity

criterion of MED in fact produces the true receiver functions for the set

of stations, m particular, if some feature is common to all receivers, it

will probably be removed from all receiver estimates. This, however, is

not considered a likely occurrence. One possible check is provided by

ccmparing the varimax norms for synthetic seismograms at all stations

produced using a reasonable source function, attenuation operror. etc.. with

the norms computed for a simple event, well recorded at all stations. If

the degrees of co.plexity, as indicated by the varimax norms, agree for

data and synthetic, then this is an indication that no large scale common

i^^ttfe^j-^a^a^'Z

■■M f.l.h ■ -^.■_ — I, .. ■— -M- —.■■—

i iiiiinj'WiillWIWWmnHWmwBBiiiiiiii

29

structure has been eliminated. Additional tests are possible for sources

where independent near field estimates of the source are available.

Q

aasHBi

■'■—■■-•■ . ' • J,l iii.iMimir.iiii •• ■ " ■-Mil,; ■ '. ' ■ ■"■ •„ ■ ....-■.■■■ - . '2

30

U

4.2 APPLICATION TO DATA

The methods of the previous section have been applied to events recorded

by the SDCS stations located at Yucca Flats, denoted by YF1 through YF4,

and a station 0B2 located at Climax Stock. In general, seismograms

recorded at the YF stations, which are located in a dipping basin, show far

greater complexity than do the same events recorded at 0B2. Events from 3

azimuthal windows (northwest, southwest and southeast) have been examined

in detail.

In order to test the deconvolution and stacking procedure, transfer

functions obtained from the procedure were used to predict the seismograms

at the YF stations from those recorded at 0B2. This also provides a

check on the assumption that near station structure varies sufficiently

slowly that variations in azimuth and ray parameter from event to event

within each azimuth window do not produce large amounts of scatter in

the receiver function estimates.

As may be seen from Figures 14, 15, and 16 there is very good agree-

ment between observed and predicted waveforms for the YF stations. While

many of the differences that do exist may be attributed to the presence

of high frequency noise at the YF stations, there are variations associated

with small changes in azimuth or ray parameter that must result from three

dimensional near receiver structure. These appear primarily as large,

incoherent arrivals late in the record. Since a large degree of interference

is present late in the record, particularly for complex sources, it ij not

surprising that deviations from an "average" transfer function become

apparent in this portion of the record.

fe-1 •m.am^m~

- . . . -. , -. - _ ... . ■'•--- "•'TirTiirTTrifrrz^i

31

•\irr.Mucf. STATIONi /io70BorsNv SO. 2«

0.00

«0

10.00

STHTHES'/fD «fre^O 123.Cl

•novoRTi'-KV NO.

10. CO

nr.'fntHcr. siniioiii moaiatiHi IS.S3

0.00 3. CO 10.00 s.cc 1C.CC

Kfr.KtiCf. STBTIOhi 772e0eflB2hV «2. 4(1

-st. is 10.00

artmiEsUFO ntcoio 7726üerr-*v NO. S fiB.e»

o. CO s.cc 10. CO

*."!. «4 Oh. 770408ae2NV NO. i

1 l\ A P A A (\ HA A ^l/vV P — i -80. 07

0. CO

nr.rf.nr.iict STSTIONI V704098B2N» ;s.i4

-31. 1» 0.00

f,.5"^fhCK STBTIOhi /70S090B2hV B. 44

-SO. ^2

-*. IS 0. CO

nrfcnchce STBTIOK. /70SOSOB2NV

17.75

STNIHtSIZfD «rCOID 7704C9TF-hV IBS.17

-ise. «s o. oo s, oo

STNTHMIZfC «tCfl« 770409Yr-h. 120.90

NO. S

coo

a^NTMf.SHFO fitCOW 770Ü09V»-NV 2!.tS

a. cc

-70. C4

10. c:

s.ca

STNTHtSÜFD «fCBIO 770605ir-NV (4.04

0.14.:

10. CO S. CO 10. CO

Figure 14. Observed selsmograms .-it stations 0B2MV (left) and YF2NV (right, dashed) I'rnm events to the northwest. Synthetic seismograms (riglit, solid) for station YF2NV produced by Convolving the OB2NV observation with the YF2 receiver functions appropriate to a northwest azimuth.

THIS PAGE IS BEST QUALITY PRAGl'ICABLB

m>M CiOüPY FUK1NI.SHED TO DpC

12

r ,

ntrrnr.hU aTRTioin noecECBiHv

o. CO

-3«..3 C. CO

S.CG

IffFflfhCf bldTJONi /V220B0ejhV 5S. VB

10. CO

iO. CO

srwiVjuro fifroio T;o«oBir-HV NS. u

0. 00

aihrHf.iifo firreio ivjaoeff-NV L42. ;9

STUTHES'.C'D flfrOIC 77:iQf,.'-KV NO. .C

0.C0

3TNT!lt3Ilifn nP'O^O 7VU4C51" NV lev..H

a. cc 10. C3

-^7.i3 o.oo S.CG ;c.oo

3TMhF:s;iro nrx.aw rmawr-hv S2.;:

NO. .2

0. CO 5.GC iC, u j

u ••crrchvoi v'B aiCNSi»» S.bV

-s.r-j o. CO ü. CO 10. cc

••;;« coisih aTftrKEn orrahvoi.uTioh.« 1.H3

o. cc

Figure 14. (Continued)

BEST QUALITY PKACÄC^iB. »BIS PAGE IS & CöirY FUÄISHBD TO DRC

a.t,ni' ^„n . :*A.i.,..J.L.it'. fl« |M|i. I

33

utrr.nr.HCt SIRüBHI noionofui^ to. n

-75. IS

"IffRrNCt 3THTISN1 ;va407or2Nv

tc. CO

-li. iG ic. r,c S. CO 10. CO

24. 10

0. DU 10.00

ärmiiEsurn nr-oic vvososvr-KV 53. a;

-G2. IV 0. CO

nr'rF"«NCi! 3Tn:i3.ii V7i309CB2NV 30.42

0. CO

NO. b

10.00

ardTHEsuro ntr.cio 77l!09Tr-NV 89.^5

0. cc

fltfrffc»i:r STHTISNI /729a7oi!2Nv 3-1.»2

NS. C

0.00

nr.rf*.'*'.? SIRIIONI /VSSOSOBSNV ev. es

io. oc

NO. 7

-97. b<)

STNTMESUfB «ECa<ID 772907Tf-NV aas. ,'7

NS. 0

0. CO

■IBH. ;2

SINTtlEsrtED HECMO 7723CS1F -NV 2oa.ne

s. ca o. CO s.co 10.00

Figure 15. Same as Figure 14 except the events are from the southwest

ami the receiver function is acliusted accordingly.

THIS PAGE IS BEST QUALITY PRAai'IC^BLE

jraOM COPY FURM1SHED TO BftC ________

:..*i^.v.';!;.:^^::-^.4:;.J..^^v.. ... ... . , ■•-^■■^i

p nr.rrnr.hcr. ainnoni nivosodsiiv IC'J. <!4

isn. ■. NO, ,C

l/yVVv'/V\AvWH

;o. cs

• ■Hf :5NV3! VSQ SIGNBI 5. CC

10, CO

• TIHf OCIHIN STflCKEC BKrOHVflL Ji !:h.. 2.20

34

4C. 4 /

ic.::

■'■.!.■■■ II =13 77J1S7TF-NV NO. 10 ssa.t;,,

.llNT'j'-IifC ".'"5 aoü.ij; _.

C6' C-NV ha.

i 1

■jej. ig | -. OS

''.l':llf:j;i,'.,3 HECS^O

^kvw ft

NO. ..: ni". ■C1. - *. 6*42

^AMf^lf^. a, :c

P>*

Figure L5;. (Continued)

~'-■..,.

35

•tirr.nr.Hct sionohi /vioosoBiNv NO.

ae. as

-«4. tS O. CO S. CO to. cc

STNVHFSWrD (IfC« 7710091»-KV ISO,.S

NO.

1Q. OC

-Jü.ra

DID 7V:;C!1« -NV HO. 4

wWi, Afk ]ß\ M 10, CD ■.a. co

-28. V7

92.02

10.00 S.CO tO. CO

nrrr.nr.hcr. STBTIONI ^let^ef^^v 97.92

0. CO 5.00

NO. 4

tO. CO

9S.ie JIG 77.eC9r«-NV na. 4

*! A ^/IAA flL IAJ\ \ \ Ä A vV-^

vvvr fi)Mj yw A^

s. cc s. cc 10. 00

O

K'f.'MKCr. ilRIIOKi ;7a00SflB2KV

-9. «10

SINTHeSUFO ffCOIC 77SCCSTr-l«V St.2B

-25. CO 0. CO

Ü

WKru'Ncr a;ar!3hi J7a70»BÄ2NV «6.411

NO. C

10.00

üTNTHfsIifd RfCOfIC 77C?0inP-KV 120.'a

-91.2» 0.00 S. 00

NO. r

10.00

o

Krr.nr.hct üSUON. V70S0S892NV 2«. es

0.00 S. CO

NO. I

10. CO

61.7« OflC 770506Tr-NV NO. ;

^^/V\/HM/ 0. 00 10.00

Figure 16. Same as Figure 14 but for events from the southeast ani receiver function is adjusted accordingly.

THIS PAGE IS BEST QUALITY PRAGHCtABLI mm COPY FLMISHED TO Bp(C

o

•seii-i mm , .-^at

36

-M.iiZ S. CO 1C.CC

aTKTHf3!irO fISrOlO VV^COCT'-HV zu«.es

-iOS.49

WyWVwVH/^ o.sc

nEfERfiiCt aiariOH' /vozofOBSNV VO. 49

-5i.41

AArWV C.DO 10. CO

3ThT!lt::UfD nf.rOlO 77C.CB1F-KV

-lib, II: D. CC 9.CC

NB. 9 rmp. ra^. - c. »se?

10. CD

nF'fl'N« StflllSNi /723CtS»JhV 40. Uf

NJ. .0

-24. .'9 0. CO S. CC 10. CO

3ihTME.i!Zfo nrcoin mswr-Hi 77. .-a

NS. .0

10. 10

MfMSHCE STBTIOKi /70S08aB2KV 20.2«

NO. kl

-lS.i;S 0. CO s.co 10. CO

S7. CS

-S7.Ü4

ma 7705oeYF-Nv NO. .1

^UUM ̂ V

fi N/U

v \ , t

s.co IC. 00

•DCCSNVOI VEO 3I0NSI.«»

3.1S

0,00 S.CO 10.00

iT'.nt OCiRlN SIRCKEO OECBKVBLUriflN«« 1.S7

-1.48 0.00 5.00 10. CO

it^0' $0

0$»

■^^

Figure 16. (Continued)

r-rr-y-;- ^""'^ V.

..,...^__ ^. ■._

37

The transfer functions obtained from the deconvclution and stacking

procedure were then processed using the MED technique in order to obtain

estimates of actual receiver functions. Receiver functions obtained in

this manner are shown in Figures 17, 18, and 19. The small, precursory

arrivals present in the YF receiver functions are largely the result of

noise in the YF records. The finite width of the pulses in the receiver

function is due to the band limited nature of the data and the presence

of high frequency noise at the YF stations.

As expected, receiver functions for the station 0B2 are considerably

simpler than those for the YF stations. In general, receiver functions

for the YF stations are characterized by three or possibly four moderately

sized negative arrivals following the first large positive arrival.

These, in turn, are consistently followed by a large positive arrival. The

early negative arrivals appear to show some moveout and amplitude

variation as we move from YF1 to YF4. These arrivals also exhibit

significant amplitude variation as a function of azimuth. This azimuthal

variation is sufficiently large that transfer functions derived for one

azimuth do not, in general, act as particularly good predictors for other

azimuths. Interpretation of these early arrivals, in light of the known

average structure, will be done in the next section.

Due to the finite width of arrivals, it is not possible to predict the

amplitude variations of incoming arrivals directly from the receiver

function maximum amplitudes. We may, however, use the receiver functions

and a synthetic incoming waveforra to predict amplitude variations. Figure 20

shows the waveforms obtained using a synthetic underground explosion source

convolved with a short period instrument and attenuation operator with t*=l. As

expected, the YF waveforms show considerably more coinplicat;ion than the

■ —.—■.. ■-■._;.„..; M

2C,'J7

YF 1 0' ^s

-23.^7

38

s,oc

o

31 35 .

YF 2 3 co

-21.33 L 0, CO b. cc

A . i\ u I \/Wl/vVi

3 00

YF 3

( ) YF 4

19 74

0.00

-19, 74 0.00 2.00 4.00 8,00

Li

OB:

IS. .IG ,

^l!i. 40 I 0. CO 4. 00 6 OC

...J

Figure 17. Receiver functions for the 4 YF array stations and

the station OB2 appropriate to a northwest azimuth.

39

I

O

YF 1 —-y ;'■' VA /VA>A-7v\rs^\ r^\

YF 2 0. DO

0 30

V/vv

JS. 4^

~n

YF 3 3. ca

-53. 43 0- cc

YF 4

33, t-.

-33. 315 0. CQ

-"vV W^/v'w'^^^^^^^

.20 4.2a

X

OR 2 1

-lb. 31 I. 3 C

Figure 18. Receiver functions for the 4 YF array stations and the station ÜB2 appropriate to a southwest, azimuth.

. ..-.. .-.M.i'n ^ ,..:.. ■.-^t^^.-_;.■,-*

40

U

YF 1

^•1. 7 7 r , z1 A

- vj v V v v

-24 77 I 4.3:

A A; J \ v'l

YF 2

-21, 7C

, v y

1

YF 3 v \ ^ ly 1 A . / I A h

-ZI 3C

L.J

YF 4

-2S. 73 4. CO

iS. 14

OB 2

Figure 19. Receiver functions for the 4' YF array stations and the station ()B2 appropriate to a southeast azimuth.

"• •—■—'^^'»—p—^——m

u

■iimiwiiiah,, -K^^J^ »»■(■.-»■■■^■■^■ii.«.!^^.,.^ ■.:-■...■^--■....

41

YF 1

r

YF 2 A ' \

-ZJCD COL

^

YF 3

-2JS; C3L

A / \ A /-. ,->

•■ ; \

Ü

YF 4

a. sä 2, ca

r~

OR 2

Figure 2()a. Syntlictic incoming P-waves computed by convolvinp receiver Functions shown in Figure 17 (northwest azimuth)

...lim «in I—^-_ ■! i il

42

©

YF 1

2302. CO

3 c: i

< ::

\ / N

YF 2

-2303 ::L

I \ I \^

s o:

U YF 3 a. co

-2 3",: c: o. CO

/ \7 v • k. CO

/ 'A

YF 4 D. CC

-2300. CO! 0. CO 2. CO 4. CO 6. CC

o OB:

J3:c

\j

o Figure 20b. c,,nt-!ietlc incominp P-waves computed by convolving the receiver functions shown ii Figure 18 (southwest azimuth) with an explosion source.

I ■IIIIIIIWIIIIHWIIUI»

-.»,*.

43

?3n?.nn

YF 1 o. o

4. 30 5 00

2300 ::

YF 2 ao:

-2303. OOL

^/

YF 3

-2300.001 4.30

o

(I

( I

YF 4

OB2

2330.00

-2300. 00 0.0

2.1C

4.00

rv

6 00

V

IBBigBBgg»

-2000, 331

4,2''

Flpure 20c. Synthetic Incoming P-waves computed by convolving the receiver functions shown in Figure 19 (southeast azimuth) with an PvnlnRinn sniirrp.

i i im ii 1'- Uta

44

0B2 waveforms. The YF receiver functions result in a factor of two ampli-

fication of the synthetic waveforms relative to 0B2. This amplification

is fairly well explained by the contrast in the shallow velocity structure

between 0B2 and the YF array. Figure 21 shows a range of waveforms and

amplitudes that result from various near-surface velocity structures.

In order to compare the ■ •rioub synthetic amplitudes, we have normal-

ized by the amplitude of the 0B2 waveform at each azimuth. The log of the

maximum amplitude for each station and azimuth, referenced to the maximum

amplitude of the YF1 station for the northwest azimuth, is shown in Table 1.

This table shows a 0.15 magnitude variation that is a function of both

azimuth and station location.

It should be noted that azimuthal amplitude variation can be caused

either by azimuthal variation at the YF stations or at 0B2. With the

present data set, there is no way of discriminating between these two

possibilities. However, within each azimuth the data show a consistent, steady

increase in amplitude from YF1 to YF4 corresponding to about .1 magnitude

units. On the basis of reciprocity, this correlates with the east-west

trend-in magnitude bias observed by Alewine (1977), as shown in Figure 22,

for nuclear events in the Yucca Flats region. The bias found in this study

is less than that discussed by Alewine. However, the YF array extends

across only the eastern portion of Yucca Flats. The bias reported here is

based strictly on reciprocity and is clearly not related to any questions

of lateral variation in near source material properties or source coupling.

As discussed in the following section, the receiver functions are the

result of complex wave propagation in the basin. The 0.1 magnitude bias

across the YF array is most probably the result of elastic wave interference

controlled by the shallow structure of the bas.in.

—-■■■: ,.■,..: """ ■ . ~

U>

©

sec 0 2 4 6 8 [ M I t I I I I

Or

5 6

km/seq

I I I 1 M I I

Or

5 6

45

SEDIMENT AMPLIFICATION

^ Y— 1|^- f1- if^ 11^ TT T I I I I I T I I I I I T-l rm Mill I ' I M ! I : I

0.2 | ,0.3 i ,0.4 ..0.5 0.5

i.8 A^- 4^- 4£- rTTTTT-m nil Tmn rr-r-n-r-m i ' i i i i i n

Figure 21. Amplification effects of shallow low velocity sediments upon crustal models. Amplitudes noted are relative to the bedrock model in upper left. Sediment thickness in kilometers Is indicated in upper models.

I.;'.^..:,:/- V:-.'; "-..-■ " .- '■■•» i——:———i—■—-—— '—«IIM»» ... II.I.I.MI .ii ji H»IUIIIU.J

===== —T:

46

Table 1

Magnitude Anomaly Associated with Receiver Functions For Yucca Flats

YF1

YF2

YF3

YF4

NW

0.0

0.05

0.08

0.10

sw SE

-0.04 -0.05

0.02 0.00

0.03 0.01

0.07 0.03

Q:

-

ast-i'.1j;gf',:...t-srr

, I III—I III

47

T—r i—r

Q

/-••■

J L J L

T.

w (U

3 ij •H

6S

E

H Z

•J; c c

•H

(N

(1) M 3

•H

'• ;:, S-i.' ,■■ •'■• ■■.

^—^«^ ^ III' ,|«IM—Mi.J^—.-..^ ^

48

4.3 INTERPRETATION OF THE YI RECEIVER FUNCTIONS

U

The shallow crustal structure of Yucca Flats has been extensively

studied by Herrin (1979) using a combination of gravity and seismic data.

Based on this study, it is possible to develop a simplified crustal model

for use in interpreting the receiver functions obtained in the previous

section. The model used is shown in Figure 23. It consists basically of

a tuff layer, dipping west, overlying a paleozoic basement. An additional,

non-dipping discontinuity is introduced in the tuff layer by the presence

of the water table. Further complications in the structure are introduced

by the presence of faults near the westernmost station, YF4.

Both the glorified optics (G.G.) method of Hong and Helmberger (1978)

and plane wave theory for simple dipping structure were used to derive

synthetic receiver functions for this structure. The results of the G.G.

calculations for one azimuth are shown in Figure 24. These results show

reasonable qualitative agreement with the early portion of the observed

receiver functions in Figure 17. The synthetics consist primarily of a

direct arrival, a P reflection from free surface and then the water table,

and a P reflection from the free surface and then the basement. Multiple

reflections and conviirted phases are small and although included in Figure

24, the amplitudes are quite small. The amplitude of the water table

and basement reflections agrees reasonably well with the average size of

early negative reflections in the data, and the moveout exhibited by the

basement reflection, relative to the water table reflection agrees

with the data.

The application of the G.O. technique to the simplified basin structure

produces a receiver function that approximates the early portion of the

l..;,wjMfe,^r>:.^:-, .-.!, . . ,.,..■....."■■

-—--.,. -~ , , ,,,,.1.

m ■jiu'k

- n «mmm. . Ill

49

I.

4J U 0 2

E

0> 0)

n

>

o Ü 3

tn o M Ü <

0)

•H >w 0 u

4J n <0

l

(0

w

f^

O «N

in

M

0)

vo fN

>1 >i >i 4J 4J 4J •H •H •H tn (0 cn ß c c 0) 0) 0) -o "O -o

Ü Ü Ü 0) 0) 0) n n tl v. V ^s

§ g

m

JQ A a AJ •• H H m 0) U M •H 0) 0) 4J *J 4J •H «3 «J Ü 2 » 0

■H 0) » 05 > 0 H > 5 H

3) O

•H « TS A 0 4J N «1 n n 0 6 4-1 IW 0

•H «M «M f*

X 0 g g 04 M a a <

CO CN

0) M 3 60

•H fa

I

50

IT YF-

i

VJ

YF-2

YF-3

n YF-4

o

Figure 24. Delta function response for the 4 YF array stations calculated with the Glorified Optics technique for the structure shewn in Figure 23,

. .;-.-.:..-v; ..:■: ^ ■—— —T" '

K 1 P w lii

1' ■r'

E

K

|

0

s

51

observational receiver functions. However several details observed in the

data are not predicted in the model. In particular, the simplified basin

model does not predict at least one early negative reflection. Nor does

this model correctly predict the timing between the first arrival and the

water table reflection. Both of these discrepancies may be explained by

the presence of a shallow sediment layer not included in the model. Some-

what more serious is the fact that the proposed model does not correctly

predict :, a large observed station to station and aLimuthal amplitude

variations of the early reflected arrivals. Nor are any of the large

arrivals following the early negative reflections predicted.

The failure of the model to account for the station to station

amplitude variations of the receiver function is reflected by the failure

to predict correct amplitude variations for synthetic seismograms

constructed in a similar manner to those in Figure 20, but using a synthetic

rather than an observed receiver function. As may be readily seen,

seismograms in Figure 25 not only fail to show the structure of those in

Figure 20, but also fail to show any significant station to station amplitude

variation.

There are several possible causes for the discrepancies between

observed and synthetic receiver functions, These include large lateral

variations in alluvium and basement depth, irregularities in the base-

ment, and systematic changes in the velocity of materials as a function of

location in the basin. Several of these factors are known to exist. For

Instance, basement maps of Yucca Flats show a large closed depression in

the basement several kilometers to the north of YFNV. Rays coming through

this depression could well have their ray paths altered sufficiently to

produce critical or near critical reflections from water table or base-

ment lavers. Similar effects may be associated with discontinuities

wM9nwirTO*»]ssitapt»

&.■

Best Available

Copy

'—-"'-'-"-—"«"

CM

LL, 0" I C J LL

52 CJ

LL.

LÜ

CJ CO

U- V—

L k S. 5 en

t j

ll.

o

LL

o rn a Li.

5 2

'igure 25. Synthetic short period P-wnve predicted by the Glorified Optics roee-iver Functions (e.g., see Figure 24) for an explosion source nt three different nzimuths (MW, SW, and SE). The seismpgrams

are normalized independently for each azimuth. No arrivals can be seen at the YF 4 station from the northwest in the Glorified Opt ics calculation.

——. '——i '-i—;—.-_■-_, ■ ^TP^STI^I^™^1^

Li

,

53

caused by faults which are known to occur in the region. In general, the

only way to differentiate between these possible causes is to fully

incorporate the detailed crustal models that are now being constructed

into the modeling procedure.

jjlv^l^:v LJ^^AMKI^ ■■.■■ ■; ■■■. ■ ■: MMfcT*"

"

^ ■

54

Ü

V. STUDIES OF OUTGOING SEISMIC ENERGY FROM YUCCA FLATS. NTS

This section discusses waveform calculations for outgoing energy from

sources located across the Yucca Flats basin and a comparison between some

of these calculations with teleseismic observations Tijeras, Tan and Commodore.

Earlier studies (Alewine, 1977) have detected an east-west magnitude variation

for events located at Yucca Flats. In Section IV we demonstrated that

the receiver functions for stations YF-1 through YF-4, computed from

incoming teleseismic P-waves, result in a magnitude bias across the array

of 0.1. It was the aim of this study to determine if the best current

estimates of the shallow structure of the basin combined with the Glorified

Optics (G.O.) technique of Hong and Helmberger (1978) could model the observed

bias.

The basin structure shown in Figure 23 was used as the basis for

computing the theoretical near source crustal response. This profile

represents, in simplified form, the shallow basin structure near the YF array.

The three-dimensional character of the basin was modeled by smoothly closing

this structure to the north and south. The closure was based on data from

stratigraphic maps of Yucca Flats. In the previous section receiver functions

calculated from this structure were not in good agreement with the observational

receiver functions. Thiiä model correctly predicts the relative amplitudes and

moveout across the array of the first reflections from the water table and

the basement complex. The calculated one way travel-time across the shallowest

layer is systematically .15-.2 sec too long. In addition, one significant

first multiple, negative polarity arrival is not predicted by the model.

Finally, a large positive arrival, arrival time = 1.2 sec after the first

P-wave is also not predicted by the model. The inadequacy of the model to

v.,- •., i*

■ •

55

O

account for the complexity of the observed receiver function suggests the

model may not describe in sufficient detail the structures that control the

magnitude bias observed at Yucca Flats. The Glorified Optics technique has

has been shown in previous tests to adequately model wave propagation in

smoothly varying three dimensional structures (Hong and Helmberger, 1978).

The technique retains information on the amplitude, timing and phase

distortion (critical reflections) of geometric rays. The technique does not

model diffracted wave propagation. Hence, to the extent that diffracted

energy is important to the complete modeling of the structure, G. 0.

will be inadequate. Within this project, discontinuous fault structures have

been modeled as slightly smoothed structures. The normal faults in the

center of the basin have been replaced with half cosine functions that deviate

from the structure shown in Figure 23 by less than a hundred meters. The

structure used in the computer program and the ray paths for the four most

significant rays are shown in Figure 26 .

In the initial stages of calculating the elastic response of the basin,

for both the vertical component for incoming teleseismic energy and outgoing

P-waves from local sources, a large family of rays was examined. For the

Yucca Flats structure, G.O. predicts that converted phases (e.g. P-waves

converted at a boundary to S-waves and then later converted back to P-waves)

are relatively unimportant. This suggests that the receiver functions

calculated from incoming teleseismic P-waves should provide a good estimate

of the outgoing wave shapes provided the function is corrected for the

effect of source depth. To a first approximation, this can be accompi-ished

by adding in the phase pP. This approximation clearly assumes that reflections

from layers between the source and the surface are negligible. The event

•

zrzr: — ■ ."j " .■'""■ ')';:■!

".f^M W*

, - • ■ ■ ■ _ _: ■- if, ii i-inini-r-Bii^iiiiiT i '-.: ■:

56

\ \

Ü

Figure'Ib. Four principal rays used in the Glorified Optics analysis of outgoing seismic energy from Yucca Flats.

60

—- .........y, ■:,l.'i„.:s»:-l^-..i. i.... Jin ■ip^i

—r—. j'.j...'"■."..i'.; ■

57

U

Tljeras was located midway between stations YF-1 and YF-2. Synthetic

waveshapes have been calculated using the receiver functions from Section IV

for the WWSSN stations COL, MAT and ARE, Figure 27. The receiver function

for the northwest azimuth from YF-1 was used for COL and the receiver function

for YF-2 was used for MAT. The ARE waveform was generated with the south-

east function for YF-1. As may be seen in Figure 27, the timing and amplitude

of the first four peaks and the approximate amplitudes and timing of later

arrivals fit quite well. This result confirms the predictions from the G.O,

calculations that for a simple or smooth structure such as exists in the

eastern portion of Yucca Flats, converted phases are not important.

An example of the waveshape calculated with G.O. for a source located

approximately at YF-1 is shown at the top of Figure 25. This synthetic

contains a von Seggern and Blandford (1972) source, a WWSSN short period

instrument response and an attenuation operator (t* = 1.0). The relative

amplitudes of the first five peaks are in fair agreement with the synthetics

calculated from the receiver function for YF 1, Figure 27, COL.

The event Tan was located about 1 km south of the station YF 4 and was

juxtaposed to the Yucca fault system. The source depth was roughly the

same as Tijeras and the emplacement media was similar. As seen in Figure 28,

the teleseismic waveforms from Tan are considerably more complex than was

observed for Tijeras. Further, this degree of complexity cannot be

obtained from any of the receiver functions from Section IV. The event

Commodore is located well north of the YF stations, on the opposite side of

the mapped Yucca fault and shows marked similarity in waveform to Tan,

Figure. 28. This argues strongly that the complex waveforms that characterize

events located in the western portion of the basin are not simply associated

with geometric rays. Calculated amplitudes and waveforms from G.O. for

t- : "—r -——- ^ ii..,. ^-—""^a

.J>

t

?•

C/2 < o2 w < ^ H M < H e

€

58

<

53

^> en oi to u —~—-. u u

.1-1 o c

"V

at three WWSSN stations from th

e NTS

event Ti

ed by convolving a

von Seggern an

d Blandford

si

ver

function.

r

o u

T3 W -H OJ tt fli > M O M OJ S (U 1= M M OJ

J2 M ft. 0

M ^ CO OJ ci) OJ > 4J > CO tfl CO S -H IS M

pu a. (ij , 0

U M -u •H a. o •|-, a •H <" « l-i J= OJ a

^ +j 4J M u o •a 4j

-C C -H OT CO 3

1^ CM

cu u 3 60

fe

^WHaawgaaBHMBww'jWMtaami

lj<::.'.,i*a.-|j.| .,....,.. : -,' -''"".'■•^'"^"«^ j

59

Q

1^

COMMODORE TAN

0/

320 m^

STN AZ

NUR

•bOrr,^

l9° ^!i^.,V>-vA^/v^

220 V li Pi

Mfyf^ cr-n 70° ,/||!i

* i.r "

, ! ^jyyM^ww>^

Ji

vl

260

'iÜflf^/A'.'v^-Vvwv^ APE

820

10.

1 300

' ' " n 160

lifMw^ I

KIP 2SS0 ^i

1280 N N l/vvvvy |j

^^1|^ HNR 259-ifp^/^Ml 520

V^ll

* 320

SHK 309°

I/VIA^ -'NAA'VV'VV SEO 3:5°

270

^llM^r^M

In^y 120

40

COL 336° 1 l/V^^^

Figure 28. Observed short period P-waves from Che .NTS events Gommodore and tan.

■■■ ■ ■ ■ ■ '■-' - ' ■ - ^ ' ■' ■

60

O

"

; :.:

sources located in the vicinity of the faults change rapidly with small

changes in position and azimuth. Figure 25. This is in conflict with

the apparent similarity of the waveshapes for events Tan and Commodore and

with the general pattern of magnitude bias discussed by Alewine (1977).

The goal of this task has been to combine the available computer code

for three-dimensional wave propagation with the velocity model for Yucca

Flats in order to investigate the origins of the magnitude bias discussed

by Alewine (1977). The general conclusion at this time is that the forward

calculations with G.O. and the basin model shown in Figure 23 cannot explain

the observed magnitude bias. At least three possible causes for this

failure exist. First, the model may be inadequate. The discussions in

Section IV show that several arrivals observed in the receiver functions are

not predicted by the model. If the magnitude bias is controlled by elastic

wave propagation in the basin, then the apparent inadequacy of the model is

very disappointing. It seems unlikely that we will have information even as

detailed as that shown in Figure 23 for Russian test sites. A second possible

explanation is that magnitude bias may be to first order the result of

coupling at the source and not strongly dependent upon the basin structure.

Material properties within the basin could grade from east to west. However,

the receiver functions calculated from the observed teleseismic P-waves predict

a 0.1 magnitude bias across the YF array that is consistent with the results

from Alewine (1977). This leads us to believe that the basin structure does

play an important role in the observed magnitude bias. As a third explanation,

G.O. does not include diffracted wave propagation effects. The attempts

to model teleseismic waveforms for sources located near the Yucca fault show

that the technique is inadequate for such complex structures and further

•m". ".' "■.—"' —■— -u^mLiiagE

o

I *'

t

61

suggests that diffraction is an important aspect of this problem. During

FY 80, Sierra Geophysics is developing a more advanced technique for wave

propagation in complex media that retains diffracted effects.

vmnHBB mm - , .. jnjmii

- ■

0

s

62

VI. AMPLITUDE STUDIES OF THE STATIONS OB2. RKON AND HNME RELATIVE TO THE

WWSSN MEAN

This section examines in detail the amplitude anomalies at the SDCS

stations HNME at Houlton, Maine; RKON at Red Lake, Ontario; and 0B2-NV

at the Climax Stock, NTS. These amplitude anomalies are determined

relative to the suite of WWSSN stations in the coterminous United States.

A complete study of amplitude anomalies associated with these WWSSN stations

is presented in the technical report SGI-R-79-011. This section will deal

only vitli the three aforementioned SCDC stations and their relationship with

the WWSSN stations in the United States.

Figure 29 shows the location of WWSSN stations in the United States

relative to the approximate physiographic provinces. Relative amplitude

anomalies were determined for these stations for seismic sources from

three separate azimuths. To the north, Russian nuclear explosions located

in five separate Russian test sites were utilized. In all, high quality

data from thirty-six Russian explosions were obtained. The events and their

locations are listed in Table 2. The amplitude measurement utilized in

this study was the peak-to-peak amplitude of the first cycle of P-wave

motion as measured on the short period vertical component of each seis.J0gram.

As explosions have theoretically spherically symetric radiation patterns.

no corrections for source radiation were necessary. To obtain amplitude

coverage in other azimuths for receivers in the United States, earthquake

sources must be utilized. However, earthquakes have an amplitude radiation

pattern due to the double couple nature of the source and may also

exhibit directivity effects due to the propagation of the fault. To iini-

mize these effects, earthquakes were sought which were simple and short-

•nrnnrnriirrnT

63

o o

o

::

x

o UJ 00

UJ

rc o u

PAC.I- ll-

cd • E W

•H H

o iw na

m oj ja

o «

01 4-1 > E

dJ 0)

0) 0) 4-1 •U C TO

•UJ -O crt OJ

Oi u 4-1 o •H rH

5 CO •H

UJ -= CN 4J CS

c c

■H C

4-i 03 TO 4J 00 W U

P co 'j: u c C/3 •

05 13 0* c u TO C

«H Z > m o co h 3 c-

UJ •H O '"

a M TO C U C tt

•H 0 4J •rl TO 0) ü >, O 4S

KJ &

01 u 3 0C

MIS PAGE IS BEST QUALITY HCÄiÜriGABM miuomFumisimmmG

'.■HmiiiimiiHiiiii nmiiw —.- —^r—

Table 2

Explosion Data Set*

-

64

£

Northern Novaya Zemlya

27 Oct 21 Oct 7 Nov

14 Oct 14 Oct 27 Sept 28 Aug 12 Sept 73 29 Aug 74 23 Aug 21 Oct

66 67 68 69 70 71 72

75 75

5:57:58 4:59:58

10:02:05 7:00:06 5:59:57 5:59:55 5:59:57 6:59:54 9:59:56 8:59:58

11:59:57

73.44N 73.37N 73.40N 73.40N 73.31N 73.39N 73.34N 73.30N 73.37N 73.37N 73.35N

54.75E 54.81E 54.86E 54.81E 55.15E 55.10E 55.08E 55.16E 55.09E 54.64E 55.08E

O

;

Southern Novaya Zemlya

27 Sept 73 2 Nov 74

18 Oct 75

Semipalatinsk East

15 Jan 65 30 Nov 69 2 Nov 72

23 Jul 73 14 Dec 73 31 May 74 4 Jul 76

6:59:58 4:59:57 8:59:56

5:59:59 3:32:57 1:26:58 1:22:58 7:46:57 3:26:57 2:56:58

Semipalatinsk East-Additional Data

5:03:00 4:57:00 2:59:00

23 Nov 76 7 Dec 76

29 May 77

70:76N 70:82N 70:84N

49.89N 49.92N 49.91N 49.99N 5O.04N 49.95N 49.91N

50.00N 49.90N 49.90N

53.87E 54.06E 53.69E

78.97E 79.00E 78.84E 78.85E 79.01E 78.84E 78.95E

79. 'JOE 78.90E 78.9CE

■ ■■^Ä^vä*;, ;W. •W*«

65

Table 2 (continued) w

Semipalatinsk West

19 Get 66 20 Apr 67 17 Oct 67 29 Sept 68 28 Jun 70 22 Mar 71 25 Apr 71 30 Dec 71 20 Feb 75

3! 4! 5; 3: 1; 4: 3: 6:

57:58 07:58 03:58 42:58 57:58 32; 32;

58 58

20:58 5:32:58

49.75N 49.74N 49.82N 49.77N 49.83N 49.74N 49.82N 49.75N 49.82N

78.03E 78.12E 78.10E 78.19E 78.25E 78.18E 78.09E 78.13E 72.08E

i

u

Kazakh

6 Dec 69 12 Dec 70 23 Dec 70

7:02:57 7:00:57 7:00:57

43.83N 43.85N 43.83N

54.78E 54.77£ 54.85L

locations and origin times from Dahlman and Israelson (1977), or Bulletin of

the International Seismological Centre.

U

o

0

^AA^&!*««fr>*tt*w mm

66

duration in character. That is, the pulses as observed on

short period records across the United States were similar in waveform,

short in duration, all had the same first motions, and were similar in

frequency content.

Table 3 lists earthquakes at a northwest azimuth. These earthquakes

occur in the Kurileislands region, in Japan, the Izu-Bonin Islands, the

Alaskan peninsula, and other places in Northeast Asia. Earthquake data

from a southeast azimuth are listed in Table 4 and are comprised of South

American events.

Earthquake data in two time periods were used in this study. Simple

earthquakes in the Kurile Islands region and in South America were obtained

for the period 1966 through 1968. A second selection of simple earthquakes

were chosen from the period September 1976 and August 1977. Table 5 lists

the events and SDCS seismograms utilized in the tie to the WWSSN stations.

Figure 30 plots the amplitude response of the WWSSN short period

instrument. Figures 31, 32, and 33 plot the amplitude responses of the SDCS

stations utilized in this study. To provide a standard basis of comparison

of these data, appropriate filters were constructed to change the instrument

response of the SDCS stations to the response of the WWSSN stations. To

Illustrate the band response of the instruments relative to sources used

in this study. Figure 34 plots the power of the Piledriver explosion

source at NTS relative to frequency. Figure 35 illustrates the effect of

filtering the SDCS seismogram at OB2-NV to have the same response as a WWSSN

short period instrument.

Three examples of the earthquake data utilized in this study are shown

in Figures 36, 37. and 38. Figure 36 plots a deep focus earthquake in Kamchatka

on April 22, 1977. The frequency content of the record at RKON is slightly

mm

TABLE 3

Earthquakes in a Northwest Azimuth'-

67

Date Origin Time Location Depth (km)

Ü

o

L

Kurile Islands Earthquakes

11/22/66

3/20/67

8/10/67

2/10/68

4/28/68

7/25/68

10/26/76

3/19/77

6:29:52.4

12:31:34.0

11:21:22.3

10:00:05.8

4:18:15.7

10:50:31.5

5:58:56

10:56:06

Japanese Earthquakes

1/1/77 11:33:42

1/5/77 22:44:57

2/18/77 20:51:26

6/12/77 8:48:05

Benin Islands Earthquakes

9/22/76 8:20:28

12/5/76 22:01:22

12/22/76 1:01:42

1/5/77 22:44:57

48.ON 146.8E

45.6N 151.4E

45.4N 150.3E

46.ON 152.3E

44.8N 174.5E

45.7N 146.7E

46.IN 159.7W

43.ON 149.OE

30.6N 137.2E

23.3N 143.8E

34.ON 143.OE

43.ON 142.3E

443

51

37

87

39

16

130

0

483

0

0

241

23.3N 142.IE 110

23.ON 140.OE 393

24.ON 145.OE 0

23.3N 143.8E 0

t

o

o

::

TABLE 3 continued

Earthquakes in a Northwest Azimuth*

68

Date Origin Time Location Depth (km) Region

Other Events

10/22/76 18:35:24 56.IN 153.3W 0

4/22/77 0:58:56 52.5N 138.8E 408

4/23/77 14:49:06 75.ON 134.9E 0

7/20/77 13:24:21 50.6S 161.9W 0

8/7/77 23:26:55 52.2N 176.2W 125

Kodiak Islands

Kamchatka

New Siberian Is.

Alaska Pen.

Andreanof Is.

*Locations Before 1975 from International Seismological Centre, after 1975 from Seismic Data Analysis Center.

*' '.

ft

TABLE 4

South American Earthquakes^

69

L

Date Origin Time Location Depth (km)

4/25/67 10:26:14.3 32. 6N 69. OW 39

Ll/15/67 21:35:51.5 28.7S 71.2W 15

2/6/68 11:19:23.1 28.5S 71. OW 23

4/21/68 9:24:35.5 23.4S 70.5W 41

4/30/68 23:51:17.9 38.4S 71.1W 40

9/30/76 8:04:11 24.2S 68. 2W 0

12/3/76 5:27:34 21. OS 69. OW 79

12/4/76 12:32:35 20.OS 69. OW 103

3/8/77 22:46:44 8. OS 63. OW ' 0

3/13/77 4:55:55 2. OS 58. OW 0

4/15/77 23:35:38 22.9S 68.8W 109

6/2/77 16:50:36 29.9S 68.6W 94

6/5/77 2:46:07 24. OS 70.5W 30

6/8/77 13:25:16 22.IS 67.3W 135

6/18/77 16:49:42 21.OS 68.7W 125

:: locations Before 1975 from International Seismological Centre, after 1975 from Seismic Data Analysis Center.

m

'mmmmm***' ■ -'— —~ •as

70

Table 5

SUMMARY OF EVENTS & SEISMOGRAMS USED IN TYING SDCS STATIONS TO WWSSN

Event

;;

■f

11/23/76

12/4/76

12/7/76

12/5/76

12/22/76 33923

12/31/76 33947

4/15/77 34657

4/22/77 34695

4/23/77 34705

5/29/77 35231

6/2/77 35206

6/5/77 35168

6/8/77 35191

6/12/77 35285

6/18/77 36493

6/21/77 J5425

7/20/77 354S3

8/7/77 35841

SDAC Seismogram Number 0B2NV RKON

34029

33789

33857

33797

33943

34659

34703

35229

35205

35169

35287

35485

HNME

34031

33859

35147

35233

35170

35487

1 il

•-- •; -r—l i l-.Jini • ■ iM^i.iilHl .ntmMi'V ^M r —

71

CL (f)

if) CO

o

3anindiAiv dV3Nii

K

! i

c

I 4J CO c

•H

O •H U 01 a

u o m

to

0) 4=

OJ co c o a co (U M

(U 13 3

a 5 CO

0)

o

(U M 3 00

■ ■ r

72

>

CM GQ O

u

BamndiAjv dV3Nn

c •H

u o 4J

ffl E

•H ■H U

T3 0)

Ü O

> 2

I (N BQ o c o

(5 4-1

co

to CJ Q C/3

01 JZ 4J

(U co c o c. CO • 0) a)

•H

13 3 4J

■U CO •H 0/ H H c e «

to 0) > f (U H Z

n cu u 3 00 •H fa

.<:■:,'.<■•.. ... ^..

——^—

73

:. -

o er

Q

i'

%

3aniiidiAi^ dV3Nn

(U

OJ ■u CO U o H

2 O

c o

u o

tu

<4-t o a) ca c o ft CO <u M

0)

3 4J •H rH • a o 6 -H

tu U

IN n cu u 3

i..!ii|-..:-.:rcr- v ■ -i.

BBfes^iifitiiiai. I'T"'

74

j

LU

.

Q

:.

aaniiidiMv dv^Nn

01 c

•H « 2

c o

3 o = 4J CO

OJ ■u CO u o

s 2 03

c o

•H *J CO 4J tn

a

(U

<4-l o 0) w c 0 c CO (U M

OJ T3 3 4J •H H a 6 CD

OJ JS H

OJ M 3 00

o

I

|

Li

m m _._.. ■ - , ,; i . . ^i. ... .. -^ -..i. i^iirtrmitfiiii

75

i

;

cr ÜJ >

Q LU

CJ

d3M0d av3Nn

(U u M 3 O m c o

•H tn o

a

OJ

> •H u -a ai

iH •H cu

Qi j: u

to c o a ca OJ u

u <u Sf o a

S ca

•H (U CO cu H (U

(1)

(U 1-1

s, 1-1 fa

o

j .■^m...a.ii'|pwiw«4t

pi

76

■i.:..■■"■■■"■'*

I

...

o

a a

"1 d 1 ID

O O

a a

o a d

a a

O

o a o

(JO o

o

a CD o

o a

to to

(U tn c o & en

C (U

I (fl c

•H

a) S n) M

a)

Ü

e « M ec o B M

•H OJ co

w o Q c/:

§ 60 C

•rl U 0) • u u

rH C •H <U y-i e

3

O 4J tn c

•H

o o

u (U

n-i en a) en

ß

O

iD C)

o d

LU

7-" LD

i'i..

or

o a □

CD CD

m

in en

(0 u 3 bO •H

s

'/)

'.'■,

IHIS PAGSB IS BEST QUALITY PRAariCABLE mM COiP^ FIMJISHED TOBPCJ

■.., V'-.]>.;,a;..?g!fff"P|N<ifi ".■*■"' "I. "

77

O

O

c •H

0) ^ ra 3 cr

J5 u M CB OJ

cn 3 U o

UJ

a. 01 Hi

T)

n)

u o

14-1

tn E « u oc o E cn

•H 0) • 'J3 r-~

r~- W a\ U H o w ^

CN 13 CM

<U ^ H 0) •H 4J U H a •H < u-

c •a 0 c CO M

C tn •H E M CO M M 3 M O

u g o W

•H ffl (U M cn u

CO Z J= CO u CO E § CO

, vO ro

0) U 3 60

«rs^1^««« X

,m~—"~°'"'- w—i man» i WWIMW «f^^^WSH *PP

,.-■■..—_ ^j-A.-Jil mmmmmmmm

78

,^

K: G

Q

O o ^ UJ ■=»

c (0

M o

4-1

B CO V4 oc • 0 r^ E n-* tn a-

•H H OJ cc 9

00 a; U aj Q c (/! 3

|-3 -a m C S-4 c (L 4-1 en H ü •H •H IU U

<u !0 e

< T3 C & en u

3 w; C u w, o a; c

•H •o

UJ d M ^ (U CO 4J 3 H O* •H -c H-l AJ

M T3 CO C QJ to

J3 tft i-'

s D. TO (U h T) 00 o (U E JJ M CO

•H •H (U •a co 0)

z g OT 0) Crt ■u P c

3 öl,

^s^™?mm

fe^j- «MM v^

■ -—-

79

o

o

o

..

c u

•H J-J r:

4J jz C 3

E O w w n 3

c a: a u

TS C o

CB -r-l

O -u U- tfl

M -U E fll re M rH OC 0) c > E 0) m H

•H o <u x m

•H M 0 o c c W CJ

13 E^ (U M 0) • iJ r^ H r~ .-i a\ U-i H

■o - a c ts i-' Cfl H C

to 0) > E C o CO 3 M '-J CO 60 O C f"

E O JJ m •H ec ^ 01 c n 03 -H U-l

M SS M (-1 OT 3 re WO u S U 3 O

to

• oo m

a» M 3 60

ft THIS PAGE IS BEST QlTAIilfY PR&an,CABJiB

MOM COPY ^paMSÜEJO lIQIJOlQ _

EffiESE3MSBZ2HESS2E: ....^^ ^UISIW»?

80

O

greater than the other stations shown in the figure. A systematic study

of the waveforms observed at RKON indicated that this station occasionally

shows anomalously different frequency content than other WWSSN or SDCS

stations used in this study. Figure.37 plots an intermediate-depth earth-

quake in South America which occurred on June 8, 1977. The event was

simple and shows similar waveshapes at stations across the United States.

Figure 38 plots a deep-focus earthquake located in Japan which occurred

on June 12, 1977.

The following series of figures plot the amplitude data for the

WWSSN and corrected SDCS stations. Figure 39 plots amplitude data obtained

from explosions at the five Russian test sites. The data at each test site

were averaged separately, and the mean for each site and each station are

plotted in Figure 39. Amplitude data for the SDCS stations were obtained

only from the Semipalatinsk east test site. Some of the WWSSN stations in

the United States are located at distances from the Russian test sites as

to be affected by the core-mantle boundary. Those stations which suffer

from the effects of diffraction at the core-mantle boundary have not been

plotted in this study or included in the analysis of relative station

amplitudes. Figure 40 plots the mean and standard error of the mean for all

of the explosion data from the Russian test sites. The vertical scale in

the plot is a logarithmic scale ranging from 1/10 of the mean to a factor of

10 greater than the mean. The stations listed horizontally are arranged in

roughly a west to east arrangement. This format is utilized in all figures

of this type in this section. The solid line in the center of the grauh

represents the geometric mean of all the stations used in the study.

Station 0B2-NV (listed as 0B2N) shows slightly greater amplitudes than the

' 'rT^,.4i.ll''','!''!1 *mi

——————

81

U

r i

I 0

: l;

i

- ^n^m^m unqAfi!*

üü^ZünJi^:: . T,"'., V'T . J '."U" ! *>nili|i!

82

0

X

t

o

en

a u cd

o^ ro 0) M 3 ti.

•H f-

c ■H

13 0)

4-1 U

0

4J Cfl

•a

aj

U-J

0

TO 0) e

14-1

o u o V-i

M cb

u CO

c CO

c CO

(U e (U

H

O

at •H

rt i i i i -ii in- HI HI -- - , :. - i r - ^. - -' - "'■ •"- ---"--" -'rrirri^iiiiiiniiffiiHmfTni - ^^m

S3

mean. RKON shows amplitudes that are nearly a factor of 5 greater than the

mean and HNME shows amplitudes roughly a factor of 2 greater than the mean.

Figure 41 plots earthquake data from the northwest azimuth. The data show

greater scatter than the explosion data, but the means for each source

region are well determined. Figure 42 plots the mean and standard error

of the mean of the data shown in Figure 41. Station 0B2-NV shows a value

slightly lower than the mean: RKON shows a value approximately a factor of 2

greater than the mean and HNME shows a value at the mean. Figure 43 plots

data from South American earthquakes. The mean and standard error of

the mean of these data are plotted in Figure 44. Station 0B2-NV shows a

value slightly greater than the mean. Table 6 shows the mean, standard

error of the mean, and the number of determinations in the calculations for

each station for a northern azimuth to the Russian test sites. It should be

noted that N in this case indicates the number of test sites utilized in the

determination of the mean. As the primary SDCS data were obtained only for

the Seraipalatinsk east site, a separate listing is indicated in Table 7 for

this test site alone. It should be noted that diffracted data is included in

this list and that the mean of these data will be different than the mean

for all of the Russian test sites that exclude the diffracted data. The

P-waves measured at the SDCS stations for this test area are not diffracted

and the relative mean amplitudes are not affected by the more distant,

diffracted arrivals. Tables 8 and 9 list the mean, standard error of the

mean, and number of data used for earthquakes located along azimuths to the

northwest and earthquakes from South America, respectively.

Before comparisons of a more quantitative nature can be considered,

effects due to differences in the geologic siting of the stations must be

evaluated. Gutenberg (1956, 1957) and Borchardt (1970) have noted that

j ■,.,,-:-..-..-.-. i ...-■.- - ■•"•'"'"-^rommfxtnnmiiw^ A —_»_______: - . ;.^^v;'V-M-:iSii;::x^.,iy.:.:^..

84

(0

3

X

DC^f- XjO dio< CM

O^ X

CI)jX> xx DP ol

X!

o

:.

<1

D<K+ jsa^© o

a

i < Kd^sa <

DCC3

@

i

■a at m

O I- c m

o u

4-1

0

C

i-

c •H -O

i- T- c

en a E 0) C Ä

M C •H

(U H (U •- 3 I- rH

^ >

6 c •H QJ

X tj

■* o

D- 0) J= > 4J -H U JJ re re a IH

E u o M -a

4J C iJ

4-1 0 re IH

•o c o a = re

4-1 .

•nan; a- en "H E < M re

•H a) co > re

•H 0) w x: w re 4J

<u o x: OS S 4-1

IU

THIS PAGE IS BEST QUALITY iWoWWE iWM CiOF-i FUKNISHEP TO DplQ

Eä*»-V

™

■ ■III !■■■■ 111. M,'.min.uiiiiiiiifii

85

'.

.

I

9

o

0 V^

M 3 U,

CO u CO

•a

u o

14-1

c

o x: uo

OJ u CO

c O) B

O

M O M P 0)

-a CO

C CO

c tfl

c CO ai B OJ

H

CM «a-

OJ !-. 3 00

•H

THIS PAGE IIS BEST QUALITY PRACItCABLB

11

—,—, :

86

-

Q

0

D

v>

<

q

s J\T vm xo<ba cp 44X\ \

< a

J».<*rT.

+

^ t\j

D

a 8

&

fe8

§ -I

1 Ü

M 9 tu

..8

«•i/1

Cft

Ut

0) >

— a»

0) u • B 01

P tn —i

c T; 3 -H er H x: c •u -x

re SJ Q) ^=

JJ C ffl I u

•H K I- C

< iJ

o AJ to —

6 3 O ^ -J

-o

•Ü OJ •H W rH JD C 0 E CO OJ

0) rH > tB

•H > 4J (0 C

iH «3

Pi s

m «a-

tu

tf.

t ~-.;y ^fc.,.-: J.j-v -,;.<■;; v-^ !.■!,■

fo *»is

7U '"""" ■; "z:— '■!"

—^'

^ 87

0

l

.

. [■

1 ,

1!

■

1 :,■■

B*'

1 5-

| jj ■

.;

;:"'-

[

i

I 1 1

Ü

i: !

1 M 0

h an e

HEH .

i '1 3H -BH

n 1—B- —1

VE —B- ^

. e B

D B

B il 1 p

O E 3

i

3 P L "

l-GH

1

o a

Ü

^

3

5

•-

ä

h 8 g

5

■a

o i—i a o u to

<r

■H Pu

E 0 u

0 c CO OJ e

u o M M OJ

-a u re

c m co

-a c CO

c (0 OJ e <u

jC H

ST

^ N ^

Qi CO

•H

I

I-.

rrrr ■ V . i- V- M

. ....:' tu i ' ir ■'■' .;

w\ 88

e

o

Table 6

Five Russian Test Sites

STATION MEAN S.E.M,

BKS 1.10 0. 14

COR 0.89 0. 10

LON 1.27 0.23

GSC 0.75 0. 11

OEl'NV 1.22

MSO 0.87 0. 16

DUG 0.54 0.09

TUC 0.66 0.02

BOZ 1.44

ALO 0.37 0.04

OOL 0.32 0. 06

LUE( 1.15 0.12

JCT 0.83 0. 08

DAL 2.23

RKON 4.89

OXF 1.19 0.08

SHA 1.44 0.15

AAM 1.4S 0.10

ATL 0.60 0.07

BLA 0.81 0.20

SCP 1.26 0.25

ÜEO 1.40 0.20

OOD 0.65 0.09

WES 0.52 0. 10

HNME 2. 14

N

4

1

1

4

2

1

1

4

The amplitude response of the SDCS station

3

4

1

HMME.

89

: ■-

ü

Table 7

Semipalitinsk East Including Additional Events

STATION MEAN S.E.M. N •—■"--"-—•- ™ ""

BKS 1.26 0.05 8

COR 3.26 0.5G 8

LON 1.52 0.06 5

GSC 0.77 0. 03 7

0B2NV 1.34 1

MSO 0.82 0.05 5

DUO 2.83 0.12 3

TUC 0.15 0.01 8

BOZ 1.47 1

ALQ 0.32 0.02 8

GOL 0.S9 0.04 7

LUB 0.67 0.05 3

JCT 0.12 0. 03 4

DAL 1.21 0.13 2

RKON 5.75 0.37 3

OXF 1.14 0. 13 3

SHA

AAM 1.90 0. 18 6

ATL 0.62 0.03 6

BLA 0.59 0. 05 •:>

SCP 1.16 0.09 5

GEO 1.69 0.21 5

OGD 0.99 0.05 8

WES 0.53 0.05 4

HNME 2.38 0.02 2

■TTW1..;;^^^--^^--

--•_ ■ —

90

i

o

Table 8

All Earthquakes to Northwest

STATION MEAN S.E.M. N

"

BKS 1.26 0. 16 13

COR 1.30 0. 17 3

LON 0.38 0. 10 7

GSC 0.76 0. 03 13

0B2NV 0.91 0. 06 7

M30 0.75 0.09 12

DUG 1. 18 0. 10 17

TUC 0.51 0.04 19

BOZ 1.24 0. 13 4

ALQ 0.57 0.05 12

GOL 0.34 0.03 12

LUB 1.02 0.09 6

JCT 1.22 0.24 4

DAL 1.42 0.23 4

RKON 1.77 0.20 6

OXF 2.09 0.18 4

SHA 1. 14 1

AAr 1.84 0.49 tr

ATL 1.14 0.11 8

BLA 1.44 0.16 8

SCP 0.95 0. 11 7

GEO 0.92 0, 15 3

. GOD 0.35. 0.22 4

WES 1.27 0. 13

HNME 1.00 1

o

I..Ml (HI, IJ.UP.1«,. |^» ' m-mit '

91

O

I i

()

Table 9

South American Earthquakes

STATION MEAN S.E.M. N

BKS 1. 09 0.08 iO

COR 1. 83 0. 13 7

LON 0.52 0.13 8

GSC 0.99 0.09 12

0B2NV 1. 39 0. 18 4

MSO 0.87 0. 10 8

DUG 1.07 0. 12 12

TUC 0.69 0.07 14

BOZ O.SS 1

ALQ 0.68 0.06 7

GOL 0.5S 0.03 6

LUB 2. 15 0.31 4

JCT 1.26 0.76 2

DAL 2.53 1.20 3

RKON 1.85 0.30 4

OXF 2. 10 0.59 4

SHA

AAM 0.81 0.22 2

ATL 0.94 0.14 11

BLA 1.11 0.23 6

SCP 0.75 0.04 5

, GEO 0.94 0.17 6

OGD 0.58 0. 12 6

WES 0.69 0.02 3

HNME 1. 18 0.00 2

tm\ —«■■iimiiiiimiiiiiM

92

yi ■■:.. ". .

hM

<

\*\

mi

^

IQ

i>-j §1

• K . i t i

\ i

! * t

:i

«

L-v 'a

o i.... ^

I

h y.

MM

|ri t ■'y

kiwi '■'fl

I

I JL

'N s-.

o U ' fc. u

1.'» i 9

0. Ö

d C

oi

ö

s 4

Q

•■■A'

N Q

13

Q R 1» -•- *• ft

8, Q

c

CM

C

u

•r-

E C

13

•H m o Q. E O g

3 ftC

THIS PAGE IS BEST QUALITY PRAiCriftLaiiB (CläuM GO* It t UMil LlifiD TO DOC ______

^^.^ii - -■ ■--'■ >■■->'■-"■'-

93

stations situated on low velocity sediments are amplified relative to nearby

stations situated on bedrock. Booth, et al. (1974) Invoked sediment ampli-

fication to explain high magnitude residuals for long period LRSM stations

along the Gulf coast. This sediment amplification effect is illustrated in

Figure 21. The receiver effect of a variety of different sediment structures

overlying bedrock are compared with simple bedrock site (upper left) using

Thompson-Haskell propagator matrices and a synthetic explosion waveform

assuming Poisson's ratio for the shear velocity and a 0.2 g/cm bedrock-

O

i

sediment density contrast. Sediment thickness of 300 meters show appreci-

able amplification. Near surface P wave velocities of 2 km/sec amplify a

factor 1.6 to 1,8 relative to bedrock; surface velocities of 3 km/sec amplify

by 1.4 to 1.5. As stations in the central United States (from RCD to AAM)

are situated on slower sedimentary materials in contrast to the hard rock

siting of the west and eastern coastal stations, some amplitude corrections

must be made. The handbook of the WWSSN contains a short description of the

local geology at each station: RCD, 2700 feet of shale, sandstone, and some

limestone; LUB, Pliestocene terrestrial deposits; JCT, Cretaceous Edwards

limestone; FLO, 50 to 60 feet of recent clay overlying Missippian bedrock;

OXF, 2100 feet of Cenozolc and Mesozoic sediments; SHA, sands and gravels

of Plio-Plelstocene age underlain by clays and sand of Miocene age; AAM,

200 feet gravel, 800 feet of shale, 4800 feet of limestone. Composing a

velocity section from the geologic descriptions is pretty much creative

guesswork. Shallow lying sandstones and shales have compressional velocities

from 1.4 to 3.3 km/sec. Limestones velocities are sensitive to the extent

of crystallization and range between 1.7 and 6.1 km/sec. Seismic refraction

surveys provide only general control as sediments are grouped in a single

layer characterized by the highest velocity arrival. Well log velocity depth

iiiii,iiiiiiiiiiiiff[tiiii#-^-"-"--: --■--■■"■-■'■'■^-^•'--.:LJ*'^^^^^^^^m'mm*-u ^^'M»l^, ' T^ÜZZIäZIÄTTTi^l

J

94

I

data provide the only accurate control, but such information was obtainable

only from SHA. The lower right velocity-depth section in Figure 21 was

simplified from a well log 25 km from SHA, and shows a factor of 1.8

amplification relative to bedrock. Given the lack of velocity control in

the surface layers at RCD, LUB, JCT, DAL, and AAM, only an approximate

sediment amplification correction of a factor of 1.4 to 1.8 may be estimated.

Figure 45 summarizes the amplitude data obtained in this study. These

data are uncorrected for the sedimentary amplification effects discussed

above. Data from three azimuths are all plotted on the same figure to

illustrate the extent of variations observed in the data. The SDCS station

at 0B2-NV shows a value that averages slightly greater than the mean value

of all data. RKON on the other hand shows values that are consistently

greater than the mean value. HNME shows values from the north which are

larger than the mean but shows values similar to the mean for earthquakes

from the northwest or south. These data are replotted in Figure 46 relative

to station location and physiographic provinces in the United States. The

symbol size indicates the amplitude factor greater or less than the mean.

Upwaid pointing triangles indicate values greater than the mean, downward

pointing triangles indicate values less than mean. The location of the

triangle relative to the point indicating the station represents the azimuth

to the events used in the study.

In summary, a collection of 36 nuclear explosions from the Soviet

Union, 22 simple earthquakes from the Kurile Islands and surrounding regions

of northeast Asia and 16 simple earthquakes in South America have been

studied to obtain relative amplitude patterns for WWSSN stations in the

United States. SDCS stations at 0B2-NV, RKON and HNME were analyzed and

tied to the amplitude patterns observed for the WWSSN stations by using a

1 l1r>liMHIIIM«lfrM lull i_

u 95

o

o

L

■•■'■'"''■■■■ - ■■ -^-J-"*--- -

96

subset of the explosion and earthquake data. The results indicate that

station 0B2-NV at th^ Nevada Test Site shows amplitudes that are systemat-

ically slightly greater than the mean amplitude in the United States for

sources in the Soviet Union or earthquakes in South America and shows

amplitudes that are slightly smaller than mean for earthquakes in north-

east Asia. RKON shows amplitudes that are consistently greater than the

mean for all azimuths, and are significantly greater by a factor of five

for explosions at the Seraipalatinsk east test site. Station HNME shows

amplitudes which are very near the mean for earthquakes from the northeast

and from South America and shows amplitudes about a factor of two greater

than the mean value observed in the United States for explosions at the

Semlpalatinsk.

Ü

wi ii i w\mm j

' A^-fc^ -■ - -..L.... ..

i im II »»i .H'.JM i

; J*. ■?«.''JLWJ*A«

97

0

References

Alewine, R. W., G. B. Young, D. L. Springer and R. W. Klepinger (1977).

Teleseismic P-wave magnitude-yield relations for well-coupled

Nevada Test Site explosions. AFTAC-TR-77-22 (SECRET).

Anderson, D. L. and R. S. Hart (1978). The Q of the earth, J. Geophys.

Res., 83, 58-69.

Anderson, D. L. and R. S. Hart (1978). Attenuation models of the earth.

Physics of the Earth and Planetary Interiors, 16, 289-306.

Anderson, D. U, H. Kanamori, R. S. Hart, and H-P Liu (1977). The earth

as a seismic absorption band, Science, 1104,

Booth, D. C., Marschall, P. D. , and J. B. Young (1974) Long and short

period P-wave amplitudes from earthquakes in the range 0 -114 ,

Geophys. J. R. Astr. Soc, 39, 523-537.^

Borcherdt, R. D (1970). Effects of local geology on ground motion near

San Francisco Bay, Bull. Seism. Soc. Am., 60, 29-61.

Butler, Ro (1979). An amplitude study of Russian nuclear events for

WWSSN stations in the United States, AFOSR Quarterly Technical

Report, Sierra Geophysics, Inc. SGI-R-79-001.

Dahlman, 0. and H. Israelson (1977). "Monitoring Underground Nuclear

Explosions", Elsevier, New York, 440.

Der, Z. A. and T. W. McElfresh (1975). Short period P-wave attenuation

along various paths in North America as determined from P-wave

spectra of the Salmon explosion, Teledyne Geotech., Report No.

SDAC-TR-75-16.

Evernden, J. F. and D. M. Clark (1970). Study of teleseismic P. II

amplitude' data, Phys. Earth Planet. Int., 4, 24-31.

mmt ■Mb ^HBa|ZBt_M rf'i^ii[iifrtor.iilHiMiHi Mfliil ■'ii ■ i 'iiiiln'MJilmlÜlifiiMiiV •itrilltt'li

" j ■! I I.III »i B J'l i HftglW'

--'---■• -A*. I II l|-H'li'l III' llrfli^ itl " 1 llf

o

98

Gutenburg, B. (1956). Effects of ground on shaking in earthquakes,

Transactions, A.G.U., 37, 757-760.

Gutenburg, ß. (1957). Effects of ground on earthquake motion. Bull. Seism.

Soc. Am., 47, 221-250.

Hadley, D. M. (1979). Seismic source functions and attenuation from local

and teleseismic observations of the NTS events Jorum and Handley,

Sierra Geophysics, Inc. SGI-R-79-002.

Hadley, D, M. and R. S. Hart (1979). Seismic studies of the Nevada Test

Site. Sierra Geophysics, Inc. SG1-R-79-003.

Hart, R. S. (1977). "The distribution of seismic velocities and attenuation

in the earth", Ph.D. Thesis, California Institute of Technology, p. 354.

Hong, T. L. and D. V. Helmberger (1978). Glorified optics and wave

propagation in non-planer structure, Bull. Seism. Soc. Amer.,

68, 1313.

Kanamori, H., and D. L. Anderson (1977). Importance of physical dispersion

in surface-wave and free oscillation problems-review. Rev. Geophys.

Space Phys., 15, 105.

Liu, H-P, D. L. Anderson, and H. Kanamori (1976). Velocity dispersion

due to anelasticity; implications for seismology and mantle composition.

Geophys. J. R. Astr. Soc., 47, 41.

Minster, J. B. (1978). Transient and impulse response of a one-dimensional

linearly attenuating medium. Part It Analytic results, Geophys. J. R.

Astr. Soc, 52, 479-501.

Peppin, W. A. (1974). "The cause of the body wave - surface wave discriminant

between earthquakes and underground nuclear explosions at near-regional

distances11, Ph.D. Thesis, University of California at Berkeley.

, . mmmmmmmmmmmmmmmmmmmm—*— ~ ' ! ' " at«liil»aMmvAi^Mia^.i^m.a»u^i»C^ ■"-"■-■W«titB^Mti-1ll»Wilinii-Aiiia>lillilf-Itiaitiihi 11 in-

'- am

99

Springer, D. L. and R. L. Kinnaman (1971). Seismic source summary for

U.S. underground nuclear explosions 1961-1970, Bull. Seism. Soc. Am.,

61, 1073.

von Seggern, D. and R. Blandford (1972). Source time functions and

spectra for underground nuclear explosions, Geoph. J. Roy. Astr.

Soc. 31, 83-98.

Wiggins, R. A. (1978). Minimum entropy deconvolution, Geoexnloration, 16,

21-35.

.,

111—iinwHiwMmi.wm