D OET Sf

RWE

AD-A203 393 YIEL ESTIMATION OF NOVAYA ZEMLYA EXLOSIONS FROM SHORT-PERIOD BODY WAVES W.W. Chain L Mc~aughlin, R.L. Cessaro, MA&Marshall, and A.C. Lees Teledyne Ucotch Alexandria Laboratories 314 Mmomoary Suvae Alexandria, Virginia 22314-1581 JUNE 1988 FINAL TECHNICAL REPORT: TASKS 1 and 2 ARPA ORDER NO: ADS5143 F ROJECr TITL: Yield Estimtion and Regional Location CONTRACT: MDA9O3-87-C-0069

Approved for Public Release; Distribution Unlimited1 Prepared for

F Sf OET D Ei

DEFENSE ADVANCED RESEARCH PROJECTS AGENCY 1400 Wilson Boulevard Arlington, VA 22209

~

Monitored by:W Defense Supply Service - Washington Room ID 245, The Pentagon Washington, D.C. 20310

BL

The views and conclusion contained in t repait ame hose of t authors and should no be interpreted as repiesenwig the official polices, eaier expressec or implied, of te Defense Advanced Research Prcoects Agency or die US. GoveinmeaL

R8

12

15

082

Unclassified

AP

SECURITY CLASSIFICATION Of THIS PAGE

REPORT DOCUMENTATION

3

Form Approved

PAGE

OMBNo 0704.0188 1Ep Date Jun30, 1986

la REPORT SECURITY CLASSIFICATION

lb RESTRICTIVE MARKINGS

Unclassified DISTRIBUTION IAVAILABILITY OF REPORT

3

2a SECURITY CLASSIFICATION AUTHORITY

Approved for Public Release; Distribution 2b DECLASSIFICATION

DOWNGRADING SCHEDULE

Unlimited

4 PERFORMING ORGANIZATION REPORT NUMBER(S)

S MONITORING ORGANIZATION REPORT NUMBER(S)

TGAL-88-03 6a NAME OF PERFORMING ORGANIZATION

6b OFFICE SYMBOL

7a NAME OF MONITORING ORGANIZATION

(if applicable)

Teledyne Geotech

Alexandria Laboratories

Defense Supply Service - Washington

6(. ADDRESS (City, State, and ZIPCode)

7b

314 Montgomery Street Alexandria, VA 22314

ADDRESS(City. State, and ZIP Code)

Room 1D245, The Pentagon Washington, D.C. 20310 8b OFFICE SYMBOL

Ba. NAME OF FUNDING /SPONSORING

(ifapplicable)

ORGANIZATION

DARPA

9. PROCUREMENT INSTRUMENT IDENTIFICATION NUMBER

NMRO

MDA903-87-C-0069 10 SOURCE OF FUNDING NUMBERS PROGRAM PROJECT TASK NO. NO. ELEMENT NO

8c. ADDRESS (City; State, and ZIP Code)

1400 Wilson Boulevard Arlington, VA 22209

WORK UNIT ACCESSION NO

11. TITLE (Include Security Classification)

Yield Estimation of Novaya Zemlya Explosions from Short-Period Body Waves 12. PERSONAL AUTHOR(S)

W.W. Chan, K.L. McLaughlin, R.K. Cessaro, M.E. Marshall, A.C. Lees 13a TYPE OF REPORT

13b TIME COVERED

Final Technical

FROM June 87

14. DATE OF REPORT (Year, Month, Day)

TOJune 88

15 PAGE COUNT

August 1988

103

16 SUPPLEMENTARY NOTATION

17

COSATI CODES FIELD I___

___

18. SUBJECT TERMS (Continue on reverse if necessary and identify by block number)

SUB-GROUP

GROUP

estimation, Novaya Zemlya, Body Waves

IYield

19. ABSTRACT (Continue on reverse if necessary and identify by block number)

-

This study investigates the characteristics of Novaya Zemlya explosions using

various body wave phases. This information will be used to calibrate the yields of these explosions. The azimuthal variation of amplitude for the Novaya Zemlya explosions as seen from WWSSN recordings indicates that there is a strong component of near-source heterogeneity due to multiple source excitation, near-source structural

heterogeneity, or source anisotropy. The systematic azimuthal variation in amplitude may be modeled with a sin(20) curve, which allows an estimate of the magnitude bias due to certain network station distributions. Several events that have an azimuthal variation in amplitude which departs from a sin(20) curve may be possible multiple explosions .

..

,

20 DISTRIBUTIONIAVAILABILITY OF ABSTRACT [X UNCLASSIFIEDUNLIMITED 0 SAME AS RPT 22a. NAME OF RESPONSIBLE INDIVIDUAL

Dr. Robert R. Blandford DO FORM 1473. 84 MAR

21. ABSTRACT SECURITY CLASSIFICATION

Q

DTIC USERS 22b TELEPHONE (Include Area Code)

22c OFFICE SYMBOL

(202) 697-7523 83 APR edition may be used until exhausted All other editions are obsolete

NMRO SECURITY CLASSIFICATION OF THIS PAGE

Unclassiffed

Unclassified SECURITY CLASSIFICATION Of THIS PAGE

(19. Continued)

'-

" Clear pP and pPcP depth phases can be observed in deconvolved teleseismic Pwave and PcP-wave source time functions from Novaya Zemlya events. With few exceptions, pP and pPcP delay times show a systematic increase with increasing size. The two largest southern Novaya Zemlya events observed (mb > 6.7) show delay times of about 0.6 sec., and most large events (6.0 < m b < 6.9), have depth phase delays between 0.35 and 0.55 sec. Relative explosion source size estimates are presented based on spectral measures of P, PcP, and Pdiff at EKA and WRA arrays. These spectral energy measurements may be correlated with the mb estimates to provide an independent calibration of magnitudes. By comparing the spectral energy levels between Novaya Zemlya and Amchi a events, it is found that the largest Novaya Zemlya event is larger than CANNI The time domain measurements of P' for NRvaya Zemlya recorded on WWSSN stations are used to provide a calibration of the m b estimates from P-waves which are quite often clipped for large events. Using a distance-amplitude correction obtained from the data set, the mb estimates for PM are computed using a generalized linear model. A different approach is to solve for the station-path effects and mb'S for PP' simultaneusly. Both magnitude estimates are well correlated with the mb estimates from the P~ aves.

Accesic';

For

NTiS

CR &I

DT!('

TAB

.

I

:Ucls......

St(

.,TY

..

I.

0

A

ie

Unclassified

FINAL REPORT

Yield Estimates of Novaya Zemlya Explosions

TGAL-88-03

SUMMARY This study investigates the characteristics of Novaya Zemlya explosions using various body wave phases. This information will be used to calibrate the yields of these explosions. The azimuthal variation of amplitude for the Novaya Zemlya explosions as seen from WWSSN recordings indicates that there is a strong component of near-source heterogeneity due to multiple source excitation, near-source structural heterogeneity, or source anisotropy. The systematic azimuthal variation in amplitude may be modeled with a sin(20) curve, which allows an estimate of the magnitude bias due to certain network station distributions. Several events that have an azimuthal variation in amplitude which departs from a sin(20) curve may be possible multiple explosions. Clear pP and pPcP depth phases can be observed in deconvolved teleseismic Pwave and PcP-wave source time functions from Novaya Zemlya events. With few exceptions, pP and pPcP delay times show a systematic increase with increasing size. The two largest southern Novaya Zemlya events observed (mb > 6.7) show delay times of about 0.6 sec., and most large events (6.0 < mb < 6.9) have depth phase delays between 0.35 and 0.55 sec. Relative explosion source size estimates are presented based on spectral measures of P, PcP, and Pdiff at EKA and WRA arrays. These spectral energy measurements may be correlated with the mb estimates to provide an independent calibration of magnitudes.

By comparing the spectral energy levels between Novaya Zemlya and

Amchitka events, it is found that the largest Novaya Zemlya event is larger than CAN-

Teledyne Geotech

June 1988

FINAL REPORT


TGAL-88-03

NIKIN. The time domain measurements of P'P for Novaya Zemlya recorded on WWSSN stations are used to provide a calibration of the mb estimates from P-waves which are quite often clipped for large events. Using a distance-amplitude correction obtained from the data set, the mb estimates for P'P"are computed using a generalized linear model. A different approach is to solve for the station-path effects and mb's for P'P' simultaneously. Both magnitude estimates are well correlated with the mb estimates from the P-waves.

Teledyne Geotech

iv

Junc 1988

FINAL REPORT


TGAL-88-03

TABLE OF CONTENTS Page SUMMARY TABLE OF CONTENTS

v

LIST OF FIGURES

vi

LIST OF TABLES

x

INTRODUCTION

1

AZIMUTHAL CHARACTERISTICS OF NOVAYA ZEMLYA P-WAVE AMPLITUDES

4

Explosive Source Directivity Multiple Events

7 31

SPECTRAL ESTIMATES OF NOVAYA ZEMLYA EXPLOSION SOURCE SIZE FROM P, PcP, and Pdiff Source Time Function Deconvolutions P- and PcP-Wave Spectral Estimates Pdiff-Wave Spectral Estimates Comparison with Amchitka Events MAGNITUDE SCALING FOR NOVAYA ZEMLYA EXPLOSIONS FROM PP' P'

Amplitude-Distance Relationship

Maximum-Likelihood Estimates of mb(P'P')

39 40 49 56 57 63 65 72

CONCLUSIONS

78

REFERENCES

80

DISTRIBUTION LIST

83

Teledyne Geotech

v

June 1988

FINAL REPORT


TGAL-88-03

LIST OF FIGURES Figure No.

Title

Page

Composite plot of P(a) amplitude versus distance for 17 Novaya Zemlya explosions recorded at WWSSN stations. The amplitude decay assumes a l/r geometrical spreading factor similar to the Veith and Clawson (1972) relationship which is plotted in dashed line.

6

2a

Plots of sin(20) least squares fit to the southern Novaya Zemlya amplitude data (solid circles) after correcting for geometrical spreading. The Y's indicate measurements within the noise level and upward arrows indicate clipped measurements.

8&9

2b

Plots of the amplitude for each station on a focal sphere for the southern Novaya Zemlya events. The size of the symbol indicates the amplitude in mp. The circles are well recorded P. amplitudes, the triangles are for stations with clipped measurements, and the diamonds are for stations with measurements within the noise level.

10 & 11

3a

Plots of sin(20) fit to the northern Novaya Zemlya amplitude data (solid circles) after correcting for geometrical spreading. The Y's indicate measurements within the noise level and upward arrows indicate clipped measurements.

12-20

3b

Plots of the amplitude for each station on a focal sphere for the northern Novaya Zemlya events. The size of the symbol indicates the amplitude in mgi. The circles are well recorded P8 amplitudes, the triangles are for stations with clipped measurements, and the diamonds are for stations with measurements within the noise level.

21-29

4a

Partitioning according to equal area quadrant (upper) and equal station distribution (lower) and the mb estimates associated with each quadrant for the northern Novaya Zemlya event of 07 Nov 68.

33

Teledyne Geotech

vi

June 1988

FINAL REPORT


TGAL-88-03

4b

Partitioning according to equal area quadrant (upper) and equal station distribution (lower) and the mb estimates associated with each quadrant for the northern Novaya Zemlya event of 19 Oct 80.

34

4c

Partitioning according to equal area quadrant (upper) and equal station distribution (lower) and the mb estimates associated with each quadrant for the northern Novaya Zemlya event of 14 Oct 70.

35

4d

Partitioning according to equal area quadrant (upper) and equal station distribution (lower) and the mb estimates associated with each quadrant for the southern Novaya Zemlya event of 18 Oct 75.

36

5a

Comparison of deconvolved P and PcP phases recorded at the array EKA from events located at the Novaya Zemlya test sites.

41

5b

Deconvolutions of P phases recorded at the array EKA from events located at the Novaya Zemlya test sites.

42

5c

Deconvolutions of PcP phases recorded at the array EKA from events located at the Novaya Zemlya test sites.

43 & 44

5d

Deconvolutions of Pdiff phases recorded at the array WRA from events located at the Novaya Zemlya test sites.

45

6

pP-P and pPcP-PcP time delays versus Lilwall and Marshall (1986) mb measured from the P and PcP deconvolutions of Figure 5. Events at the southern Novaya Zemlya test site are indicated with "stars", northern Novaya Zemlya test site events are indicated with "boxes". Measurements made on deconvolved PcP waves are indicated by open stars or boxes, while those made on deconvolved P waves are solid stars or solid boxes. With the exception of one northern event and the two large southern events, the data exhibit a trend indicating larger events have longer delay times.

48

7a

Maximum-likelihood P-wave averaged spectra for Novaya Zemlya event 73270 recorded at 19 EKA stations. The upper and lower error bars are indicated for each discrete spectral amplitude.

50

Teledyne Geotech

vii

June 1988

FINAL REPORT


TGAL-88-03

7b

Maximum-likelihood PcP-wave averaged spectra for Novaya Zemlya event 73270 recorded at 20 EKA stations. The upper and lower error bars are indicated for each discrete spectral amplitude.

51

7c

P-wave spectra for event in Fig. 7a recorded at EKA. The spectra is corrected for instrument response, attenuation, and estimated RDP. The horizontal bar indicates the bandwidth used to estimate the spectral amplitude of the phase.

52

7d

PcP-wave spectra for event in Fig. 7a recorded at EKA. The spectra is corrected for instrument response, attenuation, and estimated RDP. The horizontal bar indicates the bandwidth used to estimate the spectral amplitude of the phase.

53

8

P and PcP spectral estimates at EKA versus Lilwall and Marshall (1986) mb. The dotted line serves to separate P and PcP spectral level estimates. Events at the northern test site are identified with "boxes", the southern test site with "stars". P-wave spectral levels inferred from PcP spectral levels are identified by either open "stars" or "boxes".

55

9

P spectral levels at WRA corrected for t*=0.45 sec, and estimated RDP, versus Lilwall and Marshall (1986) mb.

58

10

Ray paths for P'P' (PKPPKP). The phase is best recorded at distance range of 40 ° to 80'. The underside reflection from the 650 discontinuity is generally denoted as the P'P' precursor.

64

11

An example of P'P' trace from the WWSSN station BOZ (A 60.60)

66

12

P'P' amplitude-distance relationship for eight events in Novaya Zemlya. These are uncorrected amplitudes measured from the analog WWSSN film chips.

67

13

Composite P'P' data scaled to mb = 6 after correcting for instrument response and gain. A rms fit to the data yields an empirical distance-amplitude relationship for P'P'

68

14

Predicted P, PKP, and P'P' peak amplitudes from WKBJ theory for a perturbed PEM (Choy and Cormier, 1983). Note the PKP caustic at 145' and the P'P' caustic at 700 .

70

Teledyne Geotech

viii

june 1988

FINAL REPORT


TGAL-88-03

15

Observed P(max)/P'P' amplitude ratios at WWSSN stations between 300 and 800, for the eight large Novaya Zemlya explosions. Only on scale P(max) and P'P' data are plotted. The ratios decrease drastically near the caustics of about 60' to 700.

71

16

Least squares fit of mb(P)'s versus mb(P'P') using (a) a priori constraints from estimated station corrections and amplitudedistance correction (left) and (b) no station and path constraints (right).

73

17

Station-path effects versus distance obtained from ML-GLM estimates of mb(P'P') using 7 northern Novaya Zemlya events. The dashed line indicates the P'P' distance-amplitude relationship from Figure 13.

77

Teledyne Geotech

ix

June 1988

FINAL REPORT


TGAL-98-03

LIST OF TABLES Table No.

Title

Page

1

Novaya Zemlya mb'S.

2

Sin(20) Fit to the Novaya Zemlya Data.

30

3

Novaya Zemlya mb by Quadrant.

37

4

Novaya Zemlya pP and pPcP Amplitudes and Delays.

46

5

Novaya Zemlya Explosion P and PcP Spectral Levels at EKA.

56

6

Novaya Zemlya Pdiff Spectral Estimates at WRA.

57

7

Log(W**) Spectral Estimates.

60

8

Comparison of Amchitka Surface waves and P-wave.

62

9

mb(P'P') Estimates Using Amplitude-Distance Corrections.

74

mb(P'P') Estimates Corrected for Station-Path Effects.

75

10

Teledyne Geotech

5

x

.June 1988

FINAL REPORT


TGAL-88-03

INTRODUCTION Novaya Zemlya, an archipelago located between the Barents Sea and Kara Sea, has been a frequent site for large nuclear explosions detonated by the Sovie' Union. The high tonnage of the explosions there may provide a constraint on the high end of the magnitude-yield relationship. However, on account of its remoteness in the Arctic, Novaya Zemlya remains one of the most poorly monitored test sites in the world. Thick and low velocity sediments (Eldholm and Talwani, 1977; Chan and Mitchell, 1985) perturb surface waves propagation.

The lack of regional station coverage has

also limited most studies of Novaya Zemlya to teleseismic data. In determining a regional magnitude scale for Novaya Zemlya, Nuttli (1988) found the available data set for Novaya Zemlya to have a limited distance and azimuthal coverage, which makes it more difficult to assess than the other test sites. The paucity of information regarding local geology adds to the difficulty in determining a magnitude scaling relationship for the region. This study undertakes a manifold approach to understanding the characteristics and yield estimation of Novaya Zemlya explosions by applying several methods to the study of various body wave phases. Azimuthal variations in body-wave amplitude are detected for some Novaya Zemlya events and may be attributed to one of more of the following effects: tectonic release, source directivity, or local structural heterogeneity.

The observed azimuthal

amplitude variation has a strong impact on the construction of a magnitude:yield calibration curve for these events. As a direct result of the biased station locations, yield estimations for the Novaya Zemlya events may be biased low, if the sin(20) variation is a proper parameterization of the amplitude variation. In the case of double events, Teledyne Geotech

June 1988

FINAL REPORT


TGAL-88-03

which may be the source function for some of the Novaya Zemlya explosions, mb estimates are further complicated due to interference effects. These effects should be taken into account in the estimation of explosive yields. In

order

to

utilize

seismic

magnitudes

to

estimate

explosive

yield,

a

magnitude:yield relationship must be established for each test site. Two of the most important contributions to systematic offsets in the mb:Log(yield) relationship are (1) the intrinsic attenuation in the mantle directly beneath the test site, and (2) the interference effects of the free-surface reflection (pP) with the direct P wave. This report presents estimates of these two effects for the Novaya Zemlya test sites. The maximum-likelihood deconvolution procedure of Shumway and Der (1985) is used to examine directly the pP phase from explosions at Novaya Zemlya test sites, and the attenuation parameter t* is determined from the observed spectrum. Estimates may then be formulated for the yield of each event, once these two primary effects on the magnitude:yield relationship have been determined. The largest Novaya Zemlya explosions clipped a substantial fraction (over 50%) of the WWSSN network as well as an unknown number of stations that report amplitudes to NEIS and ISC.

As a consequence, the magnitudes for these events are

suspect unless the data censoring effects are taken into account.

The magnitudes

presented by Chan et al. (1988) showed that these events had magnitudes as large as, or larger than, CANNIKIN (5 Mt). In order to provide further confidence in that result, we have examined other plausible means to determine relative explosion source sizes based on phases that are not clipped for the largest events. These phases provide additional independent Teledyne Geotech

measures of source size to complement 2

the traditional mb June 1988

FINAL REPORT


TGAL-88-03

measurements made using P-waves. The core phases PcP and P'P" (PKPPKP) are examined as possible alternative measures of the radiated P-wave energy from explosions at Novaya Zemlya.

We

extend the method developed by Gupta et al. (1985) to study NTS and Eastern Kazakh P-wave and P-coda spectra to PcP-wave spectra.

By referencing the PcP

spectral levels at the EKA array to the P-wave spectral levels for events where both P and PcP remain on scale, the (clipped) P-wave spectral level for large events can be inferred from their corresponding unclipped PcP-waves. When geometrical spreading is applied to the P-wave and PcP-wave spectral levels, the P-wave moments for the events can be estimated. For the high-tonnage Novaya Zemlya explosions, many of the analog recordings of P arrivals are clipped. The core reflected phase, P'P' with its easily observable onscale recordings, may prove suitable as an independent calibration of the source size. The PP' magnitude estimates may then offer an independent calibration of the mb estimates.

Teledyne Geotech

3

NUNN

June 1988

FINAL REPORT


TGAL-88-03

AZIMUTHAL CHARACTERISTICS OF NOVAYA ZEMLYA P-WAVE AMPLITUDES Long-period surface wave data from explosions have been found to contain a strong anisotropic component (Press and Archambeau, 1962; Aki, 1964; Nuttli, 1969; Hirasawa, 1971; Toksoz and Kehrer, 1972; Masse', 1981), and may be due to triggered earthquakes (Brune and Pomeroy, 1963; Aki and Tsai, 1972) or stress relaxation (Press and Archambeau, 1962; Archambeau, 1972).

Interference between the anisotropic

component and the explosion has also been demonstrated by Wallace et al. (1983) using long period body wave data from Pahute Mesa events. For short-period body wave, Lay et al. (1984) also found azimuthal variation of amplitude for the Pahute Mesa events, which they attributed to tectonic release or local heterogeneity in the structure. As has been demonstrated by Bache (1976), Murphy et al. (1983), and Wallace et al. (1983), the level of the anisotropic component may bias the mb estimations due to its interference effect on the explosive-generated P waves. This finding is important for the estimation of energy from explosions in Novaya Zemlya, which are known to be associated with strong azimuthal variation (Burger et al., 1986). These variations may be interpreted in terms of near-source structural inhomogeneity or tectonic release. In order to examine the azimuthal dependence of Novaya Zemlya explosive Pwaves, we studied the variation of the "a" phase amplitudes for 17 of these events. The MLE-GLM magnitudes estimates (Chan et al., 1988) for these events are listed in Table 1 with those from Lilwall and Marshall (1986).

Teledyne Geotech

4

June 1988

FINAL REPORT


TGAL-8-03

Table 1. Novaya Zemlya mb's MLE-GLM

Date

mb(a)

07

mb(b)

Southern 27 Sep 73 27 Oct 73 02 Nov 74 18 Oct 75

5.18 6.65 6.50 6.23

0.05 0.05 0.05 0.05

5.47 6.88 6.80 6.52

Northern 27 Oct 66 21 Oct 67 07 Nov 68 14 Oct 69 14 Oct 70 27 Sep 71 28 Aug 72 12 Sep 73 29 Aug 74 23 Aug 75 21 Oct 75 20 Oct 76 01 Sep 77 10 Aug 78 11 Oct 80 01 Oct 81 18 Aug 83 25 Oct 84

6.07 5.40 5.60 5.76 6.43 6.27 5.99 6.37 6.13 6.12 6.11 4.03 5.09 5.39 5.19 5.23 5.33 5.16

0.04 0.04 0.04 0.04 0.04 0.05 0.05 0.05 0.05 0.05 0.05 0.04 0.06 0.04 0.05 0.05 0.05 0.06

6.31 5.60 5.85 5.96 6.65 6.49 6.25 6.70 6.40 6.38 6.35 4.35 5.43 5.63 5.45 5.49 5.53 5.43

Given the strong

Lilwall and Marshall(1986)

mb(max)

0

mb(b)

0

0.05 0.05 0.05 0.05

5.72 7.10 7.02 6.84

0.05 0.05 0.05 0.05

5.83 6.90 6.75 6.70

0.03 0.03 0.02 0.02

0.04 0.04 0.04 0.04 0.04 0.05 0.05 0.05 0.05 0.05 0.05 0.04 0.06 0.04 0.05 0.05 0.05 0.06

6.45 5.77 6.04 6.13 6.83 6.64 6.38 6.79 6.58 6.51 6.56 4.66 5.57 5.86 5.66 5.65 5.71 5.61

0.04 0.04 0.04 0.04 0.04 0.05 0.05 0.05 0.05 0.05 0.05 0.04 0.06 0.04 0.05 0.05 0.05 0.06

6.47 5.99 6.11 6.18 6.77 6.63 6.46 6.96 6.54 6.55 6.59 4.89 5.71 6.04 5.80 5.91 5.84

0.03 0.03 0.02 0.03 0.03 0.02 0.02 0.03 0.02 0.02 0.02 0.03 0.02 0.02 0.02 0.02 0.02

-

-

y

component of apparent tectonic

release associated with the

Novaya Zemlya explosions, we seek to understand the possible correlation between the azimuthal amplitude variation and tectonic release taking the approach of Lay et al. (1984). We also investigate its relationship to various other near-source parameters, including local structural heterogeneities, double event, and focusing effects.

We con-

struct an amplitude-distance relationship for Novaya Zemlya data from the amplitude data, as shown in Figure 1, that resembles the more global model of Veith and Clawson (1972).

The anomalous low amplitude values are probably miscalibrated data.

Teledyne Geotech

5

June 1988

10

4

103

"

0 0

.6o.

0.. 04C

0 Q.. ,os

*to.

1'-j

a0

ooi

0.

0

6

16

,V

plot of

at WW

reore

0

•66a

50

40

Copst

Figure* 1.

exlsin

6

DitacTh 50

sttin.

60

0

70

60

e

P)apiu

S

40

su

6

disane

(Adea 0 )

fo

in

dahe

Fiur

line

1

omostepotofPa)amliue

exlsosrcre0t)S

........

Nsain.T

eru

atrsmlr

oteVihadCasn(97)rltosi

plte6i

ie

6he

-

- ,, ,

k iuu d

i

l

lliil

'i~I

-I

,

/

emti

60

90

66 ditnefr1

eamltd

10 pedn

90

0

7NoaaZe-

sue omliud

Disa0e plte

so

6

0

30

00

#

0oaaZmy

ea

sue

/

emti

hc

i

-

s'

-I

FINAL REPORT


TGAL-88-03

Explosive Source Directivity The initial "a" phase amplitude data are read from analog film chips for 17 Novaya Zemlya explosions. These amplitudes are corrected for the simple geometrical spreading relationship used in Langston and Helmberger (1975). The corrected amplitude data are plotted both against azimuth (Figures 2a and 3a) and on a focal sphere (Figures 2b and 3b) for 4 southern and 18 northern Novaya Zemlya events. From Figures 2a and 3a, these data display a systematic azimuthal amplitude variation that may not be attributed simply to scatter of the data. The corresponding plots of the amplitudes on the focal sphere as shown in Figures 2b and 3b do not show any systematic azimuthal and distance relationship. Following Lay et al. (1984), we apply a sin(20) rms fit to each individual set of censored data where each curve may be parameterized

by y = A (1 + x x sin(2( a - 0 ))) where A is the amplitude of the modulation in millimicrons, 0 is the phase of the modulation in degrees and alpha is the station azimuth. The amplitude, phase, and statistics of each best fitting curve are listed in Table 2. The sin(20) term appears to fit the data with good confidence level for most of the events, as indicated in Table 2.

Teledyne Geotech

7

June 1988

027sep73 ~10' 10

S10

1

-

- 2

°,

-

"" Y" -----

o a---Z__ S0

P,4

•Y

OY

10-3 Y

u 10-43 0

200 100 A zimuth (0)

300

(a)

027oct73 :: 1 0

° ++ ++

i0 S10

-3 2

"

10-3

10n-4 0

200 100 Azimuth (°)

300

(b)

Figure 2a. Plots of sin(20) least squares fit to the southern Novaya Zemlya amplitude data (solid circles) after correcting for geometrical spreading. The Y's indicate measurements within the noise level and upward arrows indicate clipped measurements. 8

O2mo'u 74

-~

S1021

0

100

200

300

A zimuth (C)

18oct75

-Y

100

10-

0

100 200 A zimuth (0 (d) Figure 2a. (cont'd) 9

300

27sep73

27oct73

. ..

to-' n4-km O 10-3 m.l-kra

Figure 2b. Plots of the amplitude for each station on a focal sphere for the southern Novaya Zemlya events. The size of the symbol indicates the amplitude in mr. The circles are well recorded Pa amplitudes, the triangles are for stations with clipped measurements, and the diamonds are for stations with measurements within the noise level.

10

02nov74

18oct75

... . Figure 2b. (cont'd)

. ..

42

27oct66

1 O' z3-

10w'

S10-1

+

R+

0

100

200

300

(a)

21 oct67 S10

S10-3S.

0

100

200

A zimmth

300

( )

(b) Figure 3a. Plots of sin(20) fit to the northern Novaya Zemnlya amplitude data (solid circles) after correcting for geometrical spreading. The Y's indicate measurements within the noise level and upward arrows indicate clipped measurements. 12

42

0 7rov68

S10-

6

+

S

S10-

100

0

300

200

A zimmth

(0)

(C)

14oct69 S1001 10-2S 0y

%

~

~10-4

0

100

200

A zimumth (d) Figure 3a. (cont'd) 13

300 c)

W7N

14oct70

loo

S10-3* 10-4

200

100

0

300

Azimuth(0

o

(e)

42

27sep 71 20 a

00

:~10-2

0 C)A

100

200

zimuth(0 Figure 3a. (cont'd) 14

300

28aug72

100

I

10

- 1

_

I no

"°,

°

,.

**

4

g

- •

10-2

10-3 10-4

300

200

100

0

+

+

~

Azi-muth (°) 0

(g)

12sepT3

"

I10-

+O

0o

10-4

•

101

~

0

200 100 Azimuth (4(h) Figure 3a. (cont'd)

300 )

42

C,

29ug 74

10l

0

01

00

0

100

200

A z'mmth

100

23ug

300

(0

75

loo

10-3

1 0-4 0

100 200 Azimuth (0 Figure 3a. (cont'd) 16

300

zL100

10-2

10-3

10-40 0

300

200

100

21oct75 100

0

10-1

0100 C~~)0Azimuth

~

117

200

(0)

300

01sep77

~0

d 10-3 10-4

200 100 Azimuth (0

0

(in)

~

aug78

10~18

Q) 10-1 10-2

S10-3

C)

0

100A

200u30 (n)

Figure 3a. (cont'd) 18

300

4

11 oct8O

loo 6

-

10 2 10-32 •S10

+

1020

S

0

10-4

.

300i

200

100

q0

Azimuth

(0)

(o)

4201oct81 0)

1 01.. 10-

1

_ ly

• yy

y

10-3 0)0

,

10I-4

0

100 200 Azimuth (0) (P) Figure 3a. (cont'd) 19

300

42

1 8aug83 0100

S10-1 10-2

10~-3 10-

0

100 200 A zimuth (C

300A

(q)

4

25oct84

10-10 10-2

S10-3

10-

0 0~

~

100 Aziuth

200

Figure 3a. (cont'd) 20

(0

300

27oct66

2 1 o.c

nw-

t

i

.... ....

in0-

o

o

:10-2 Mg_-0M o

1O'4 mR-km

Figure 3b. Plots of the amplitude for each station on a focal sphere for the northern Novaya Zemlya events. The size of the symbol indicates the amplitude in mg. The circles are well recorded Pa amplitudes, the triangles are for stations with clipped measurements, and the diamonds are for stations with measurements within the noise level.

21

07nov65

14oct69

Figure 3b. (cont'd)

22

14oct70

6d

27sep71

Figure 3b. (cont'd)

23

-~

- - -

D

28aug72

12sep73

Figure 3b. (cont'd)

24

29aug74..

23aug75

Figure 3b. (cont'd)

25

20oct76

Figure, 3b. (cont'd)

26

01

sep77

18aug76

Figure 3b. (cont'd)

27

11oc

01

t80

oct81

Figure 3b. (cont'd)

28

18aug83

25oct84

Figure 3b. (cont'd)

29

w7

7.

FINAL REPORT


TGAL-88-03

Table 2. Sin(20) Fit to the Novaya Zemlya Data Amplitude Phase 0 Date Day of year mb (ii) (degree) Southern 27 Sep 73 270 5.18 1.40 13.2 2.09 -24.9 27 Oct 73 300 6.65 02 Nov 74 306 6.50 2.63 -13.8 6.4 291 6.23 2.24 18 Oct 75 Northern 27 Oct 66 21 Oct 67 07 Nov 68 14 Oct 69 14 Oct 70 27 Sep 71 28 Aug 72 12 Sep 73 29 Aug 74 23 Aug 75 21 Oct 75 20 Oct 76 01 Sep 77 10 Aug 78 11 Oct 80 01 Oct 81 18 Aug 83 25 Oct 84

275 294 312 287 287 270 241 255 241 235 294 294 244 222 285 274 230 299

6.07 5.40 5.60 5.77 6.44 6.27 5.99 6.37 6.13 6.12 6.11 4.03 5.09 5.39 5.19 5.23 5.33 5.16

1.51 1.70 1.72 1.89 2.12 2.12 1.80 1.31 2.03 1.78 2.32 1.26 2.56 1.67 1.55 2.42 2.30 1.84

-1.4 17.2 10.4 5.1 8.2 6.2 -12.2 5.6 -4.8 -3.7 4.6 -24.3 6.7 11.3 220.9 35.3 13.5 14.9

F-ratio 1.45t 3.12t 6.22 5.40§ 3.39 6.55 7.01 7.46 7.71 5.77 3.81§ 0.33f 4.85 2.39 4.18 0.06: 4.93§ 6.04 2.61f: 12.33 11.27 3.68:

§ Fails F-statistic test at 99th percentile. t Fails F-statistic test at 95th percentile. Most of the phase estimates (0) from the best fitting curves fall within the range between 5' and 150 for the northern Novaya Zemlya test site. The same conclusion may not be drawn for the southern test site due to a lack of data. For a spherical symmetric Earth, the observed azimuthal pattern indicates that there may be a component of near-source effects or varying attenuation along the paths with different azimuthal distribution. The choice of a sin(20) fit is arbitrary and may be assessed if the azimuth

Teledyne Geotech

30

June 1988

FINAL REPORT


TGAL-88-03

amplitude pattern pertains to near-source effects, eg. tectonics release (Lay et al., 1984). The phase and amplitude of the sin(20) modeling also bear no apparent rela-

tionship to the size of the explosions.

The sensitivity of an mb estimate bias that results from a symmetric sin(20) amplitude variation are expected to be reduced if a well-distributed network is used because the nodes would tend to cancel. However, in the case of Novaya Zemlya, as shown in Figures 2b and 3b, the receiving stations are biased towards North America and Europe, and little data are available within the 0' to 800 range. If the sin(20) model is appropriate to explain the observed azimuthally dependent amplitude variation of the Novaya Zemlya data, it would predict a local amplitude maximum in this azimuth range. Consequently, mb estimates obtained for Novaya Zemlya with the present station configuration would tend to be biased low. Multiple Events Although most of the events may be explained by the sin(20) model quite well statistically, as is shown in Figures 2a and 3a and Table 2, there are a few exceptions. One of them is the 18 Oct, 1975 event in southern Novaya Zemlya which was described as a double event by Hurley (1977). Although the sin(20) fit to the data without censoring information falls within a 95% confidence interval, if the true amplitude of the censoring information were included, the amplitudes for this event between azimuths of 100" and 170" would then appear to be biased low. This effect may be caused by the simultaneous detonation of two spatially separated explosions which would exhibit constructive and destructive interference at certain azimuths. The other

Teledyne Geotech

31

June 1988

FINAL REPORT


TGAL-88-03

events that show a possible deviation from the sin(20) model are 7 Nov 1968, 14 Oct 1969, and 27 Sep 1971. The 11 Oct 1980 event, which Lilwall and Marshall (1986) indicate may also be a double event, does not show azimuthal characteristics similar to those of the 18 Oct 1975 event. We have performed azimuthal partitioning of the first motion amplitudes that are used in the maximum-likelihood estimation of the mb's for four Novaya Zemlya events. The azimuthal pattern is presumed to be due to local geological structure or tectonic release and an attempt is made to model it by a sin(20) symmetric function. Therefore, when there is an uniform azimuthal station coverage, the azimuthal effects on the mb would be averaged with accordant bias reduction. In the case of a double event, the symmetry is not preserved due to the interference effect. The partitioning is designed to demonstrate the azimuthal dependence of mb as shown in Figure 4. The quadrants are divided into azimuthal range of 0O-90 ° , 900-180 ° , 180°-270', 270'-360' (Figure 4a) and 3140-74', 740-134

,

1340-254 ° , 254°-3140 (Figure 4b). The ampli-

tude readings for the stations within each quadrant are grouped together and input to the MLEGLM estimation (Chan et al., 1988) to arrive at mb estimate for each quadrant. The mb values for each quadrant are shown in each figure and are listed in Table 3.

Teledyne Geotech

32

June 1988

07 NOV 68

0.086

55.943±

5.75± , 0.5 15.

2700

A•

±0.08

AA

05±

5.8070.056

5.571. 0.052

(a)

Figure 4a. Partitioning according to equal area quadrant (upper) and equal station distribution (lower) and the mb estimates associated with each quadrant for the northern Novaya Zemlya event of 07 nov 68. 33

I

5.260±+0.0545.9-009

2700

900

80

505± .07± .5

.16

5..259f

3144

0.056

.6

-,'

" ,

*.:.,.

*, :*: *

'

MG . .,I

-

.

. :

:

i

° '"

-

'"

"

•

°..

"

14 OCT 70 0.

6.390± 0.052

'.

.

6.423± 0.091

290

34± 0.056

S6.5

6.356±1 0,060

A

180A

6.373± 0.056

0

6.537± 0.065

•

...

i It-- or,.,ir.

l

.ail

i

'l

6

AA

li

i

~a I

m

i

ii

0,060

-

-

1

6.342± 0.055 (c)

Figure 4c, Partitioning according to equal area quadrant (upper) and equal station distribution (lower) and the mb estimates associated with each quadrant for the northern Novaya Zemlya event of 14 oct 70.

35

i"A

"

18 ocr 75

A A A6.212±

6.040±+0.073

A

A

0.116

A

A

A

AA

6.293± 0.078

6.457± 0.078

6.254± 0.083 Aa

5.942± 0.08674

0.086

d,6.447± 2540

6.292± 0.075

Md

Figure 4d. Partitioning according to equal area quadrant (upper) and equal station dis-

tribution (lower) and the mb, estimates associated with each quadrant for the southern Novaya Zemlya event of 18 oct 75. 36

FINAL REPORT

TGAL-88-03


Table 3a. Novaya Zemlya mb by Quadrant Mb

Date

00-900

900- 1800

1800-2700

2700-3600

07 Nov 68

5.943 +0.086

5.807 +0.056

5.759 +0.051

5.525 +0.049

14 Oct 70

6.423 +0.091 5.119 +0.099 6.212 +0.116

6.365 --0.060 5.156 ±0.065 6.457 ±0.078

6.534 +0.056 5.097 ±0.058 6.293 ±0.078

6.390 +0.052 5.260 ±0.054 6.040 +0.073

11 Oct 80 18 Oct 75

Table 3b. Novaya Zemlya mb by Quadrant Mb Date

3140-740

740-1340

1340-2540

2540-3140

07 Nov 68

5.571 -0.052 6.373 +0.056 5.255 +0.055 6.254 +0.083

5.900 +0.060 6.508 ±0.066 5.190 ±0.075 6.447 ±0.086

5.763 ±0.050 6.342 +0.055 5.159 ±0.056 6.292 +0.075

5.566 ±0.062 6.537 ±0.056 5.015 ±0.076 5.942 ±0.086

14 Oct 70 11 Oct 80 18 Oct 75

The variations in mb's for events 11 Oct 80 and 14 Oct 70 are about 0.15 mb unit whereas for events 07 Nov 68 and 18 Oct 75, the variations are up to 0.4 mb unit from the 900 quadrant partition. Similar variation is observed using the optimal partitioning taking into account station distribution. The quadrant mb'S for 14 Oct 70 display a symmetry that roughly follows the sin(20) distribution, as is shown in Figure 3a(c) for this event. For azimuths 900 to 1800 and 2700 to 360', we obtain a lower mb than for the other two quadrants. Since quadrant 00 to 90' is not well covered by seismic stations, the mb estimates within this quadrant may be as high as that of the opposite quadrant if the sin(20) model is appropriate. The same speculation may be made for the

Teledyne Geotech

37

June 1988

FINAL REPORT


TGAL-88-03

11 Oct 80 event due to its low yield. An asymmetric mb pattern is observed in the case of the double event, 18 Oct 75. The mb estimates for quadrants 90' to 1800 and 180' to 270' are high, as expected from the amplitude-azimuth relationship obtained in Figure 2a(c). The extreme amplitude difference is observed for opposing quadrants of 90' to 1800 and 270' to 360' as well as 740 to 134 ° and 2540 to 314' , where the variations are up to 0.5 mb unit. If the data observed for 18 Oct 75 indeed consist of observations of two events detonated simultaneously, the analysis here indicates that they were separated in space in a NWSE direction, as the quadrant mb estimates are the most biased in this direction. A similar asymmetric pattern is observed for the event on 07 Nov 68, where the quadrant mb bias is up to 0.3 unit. It is speculated that this may have been a double event separated in a NW-SE direction also.

Teledyne Geotech

38

June 1988

'.

_ , ].; L . -

_. :]

.

iI

-

FINAL REPORT

-


TGAL-88-03

SPECTRAL ESTIMATES OF NOVAYA ZEMLYA EXPLOSION SOURCE SIZE FROM P, PcP, and Pdin The attenuation of body waves from Novaya Zemlya was studied by Der et a/. (1985) as part of a study of P-wave spectra from several test sites. They concluded that the attenuation parameterized by t* did not differ significantly from the Eastern Kazakh or Amchitka test sites. Given that Novaya Zemlya is a high Q test site then near-source attenuation adjustments must first be made to properly compare tests at NTS to Novaya Zemlya. The largest Novaya Zemlya explosions have maximum-likelihood magnitudes that exceed the magnitude of CANNIKIN ( mb 6.6, 5 Mt). Because MILROW and CANNIKIN were the largest U.S. shots detonated away from NTS in what is considered a high Q region,

the

Amchitka

magnitude:yield determination.

shots have

assumed

a great

importance

for

Along with the megaton range events at NTS, they

constitute a reference set of explosions for determining the yields of Novaya Zemlya explosions.

All three Amchitka explosions, including CANNIKIN, had distinct and

late depth phases (Bakun and Johnson, 1973; Douglas et al., 1987). The strong P+pP interference for these Amchitka shots makes understanding the P+pP interference of the Novaya Zemlya explosions of prime importance when making attenuation comparisons between these events as well as those at NTS.

Teledyne Geotech

39

June 1988

FINAL REPORT


TGAL-88-03

Source Time Function Deconvolutions The Shumway-Der deconvolution procedure (Shumway and Der, 1985; Der et al., 1987b) has been used to study the explosion P-wave source function at NTS, Eastern Kazakh, and Novaya Zemlya as well as other test sites. The algorithm provides a maximum-likelihood estimate of the equivalent teleseismic explosion reduced displacement potential (RDP) corrected for attenuation, instrument response, and variations in site response across an array. Details of the deconvolutions of the explosions are discussed in Der et al. (1987a). We present results for Novaya Zemlya explosions deconvolved at two arrays, EKA and WRA. The delays for pP and pPcP as well as the relative pP/P amplitudes are measured from the deconvolved seismograms. These characteristics are intended for use in formulating a magnitude:yield relationship for the Novaya Zemlya test sites. Some examples of deconvolutions of P- and PcP-waves from Novaya Zemlya events recorded at the array EKA (A = 290) are shown in Figures 5a through 5d. A causal attenuation operator (Azimi et al.,

1968) with a t* of 0.23 sec has been

removed from the waveforms along with the instrument response and estimates of the individual site responses (see Der et al., 1987a). Measured pP-P and pPcP-PcP delays and relative amplitudes from these maximum-likelihood multichannel deconvolutions derived from recordings at the EKA array are listed in Table 4. The events located at the southern Novaya Zemlya test site are indicated by "(S)" in Table 4. The 1975291 event, regarded by Lilwall and Marshall (1986) as a double event, was clipped for both the P and PcP arrivals at EKA.

Teledyne Geotech

40

June 1988

cl

C-))

Cd

o

>~ 0

GO

U>

41) C)

4--

> 000 N

>

C:

>U

00 "CU

00

94

+

41

u94

Deconvolved Source Functions Novaya Zemlya Events Recorded at EKA

?

1983268

*

19S2284

4

(Double Event)

7

1980285

4

1979291

1979267

1978270

4

:

4

1977244

1976294

1973270 4.0 SEC Figure 5b. Deconvolutions of P phases recorded at the array EKA from events located at the Novaya Zemlya test sites. 42

Deconvolved PcP Source Functions Novaya Zernlya Events Recorded at EKA

(Southern Event)

1973270

1968312

1967294

1966300 4.0 SEC Figure 5c. Deconvolutions of PcP phases recorded at the array EKA from events located at the Novaya Zemlya test sites. 43

Deconvolved PcP Source Functions Novaya Zemlya Events Recorded at EKA ~1984299

1979291

I4

1979267 4

1978270

1976273

1975294

1975235

4(

Evcnt)

1974306

(Southern Event)

1973300

S(Southern 4-

4.0 SEC Figure 5c. (cont'd) 44

CYo 00

C-))

0

1-. 4-1-

0\0

ON

0)%-

0\

0I

co G)

I

+

0

~-400

-00 O

0

>0)M

45

:

00

FINAL REPORT

Yield Estimates or Novaya Zemlya Explosions

TGAL-88-03

Table 4. Novaya Zemlya pP and pPcP Amplitudes and Delays mb

Date 1966300 1967294 1968312 1969287 1970287 1971270 1972241 1973255 1973270(S) 1973270(S) 1973300(S) 1974306(S) 1975235 1975291(S-D) 1975294 1976273 1976294 1977244 1978270 1978270 1979267 1979267 1982284 1983268 1984299

Teledyne Geotech

Lilwall and Marshall (1986) 6.47 5.99 6.11 6.18 6.77 6.63 6.46 6.96 5.84 5.84 6.89 6.73 6.55 6.69 6.59 5.77 4.89 5.71 5.68 5.68 5.80 5.80 5.52 5.71 5.90

Relative Amplitude -0.7 -0.75 -1.1 -1.2 -0.75 -0.75 -0.7 -0.7 -1.45 -1.8 -0.65 -0.75 -1.0

Delay (sec) 0.35 0.30 0.40 0.35 0.40 0.40 0.37 0.45 0.40 0.35 0.62 0.62 0.43

-0.4 -0.9 -1.05 -0.5 -0.85 -0.55 -0.65 -0.27 -0.6 -0.4 -1.9

0.40 0.37 0.2 0.25 0.33 0.25 0.30 0.50 0.60 0.90 0.32

46

Phase PcP PcP PcP PcP PcP PcP PcP PcP P PcP PcP PcP PcP P&PcP clipped PcP PcP P P P P P PcP(?) P P PcP

June 1988

FINAL REPORT


TGAL-88-03

The time delays are plotted with respect to the Lilwall and Marshall mb'S in Figure 6.

Except for the two largest southern events, the data show a trend for the depth

phase time delays to increase gradually with increasing event size. The two largest southern events stand out from the population with depth phase delays of 0.6 sec. The time resolution of the deconvolution procedure is estimated to be about ±0.15 sec. This is the half-width-half-maximum of the Shumway-Der resolution kernel. This indicates that the depth phase time delay may be a weak function of the source size and not wholly dependent on the burial depth of the explosion unless the biggest two events were overburied.

Teledyne Geotech

47

June 1988

pP and pPcP Delay Times from Deconvolutions of Novaya Zemlya Events Recorded at EKA

(t* & instrument corrections)

7

0.

1

0.6-

1

, t

I,

I

,

I

Resolution (0 15 sec)

-

0.5-

-

0

~- 0.4-*

0

00

0 0_ 03

r-

0 N

0.3-

E0

0 0

O

0.2 oUENNZ pP

U

0 NNZ pPcP * SNZ pP SNZ pPcP

0.1

0 .0

I

5.0

I

I

I

5.5

I

6.0

I

I

I

6.5

mb (Marshall et al. 1984)

Figure 6. pP-P and pPcP-PcP time delays versus Lilwall and Marshall (1986) mb measured from the P and PcP deconvolutions of Figure 5. Events at the southern Novaya Zemlya test site are indicated with "stars", northern Novaya Zemlya test site events are indicated with "boxes". Measurements made on deconvolved PcP waves are indicated by open stars or boxes, while those made on deconvolved P waves are solid stars or solid boxes. With the exception of one northern event and the two large southern events, the data exhibit a trend indicating larger events have longer delay times.

48

7.0

FINAL REPORT


TGAL-88-03

P- and PcP-wave Spectral Estimates Gupta et al. (1985) used the average spectral level of the P-wave and P-coda for NTS and Eastern Kazakh explosions at a single station, NAO, to derive an explosion source size estimator. They found that the single station spectral estimator had an excellent correlation with ISC mb and known yields for NTS explosions. The source size estimate is made from a log-average of the spectral amplitude in a selected bandwidth after corrections have been made for instrument, attenuation, and a von Seggern and Blandford (1972) source spectral shape. These corrections pre-whiten the spectral amplitude, and if a bandwidth is chosen that has good signal-to-noise ratio, then the average spectral level is an estimate of the far-field explosion size. This level can be corrected for geometrical spreading and the free-surface effect at the receiver to yield an estimate of the explosion RDP or moment, T_,, or M0 respectively. This procedure was applied to P and PcP data recorded at the EKA array. A frequency independent t* of 0.23 seconds was assumed as an attenuation correction, and the Bache (1982) mb:Log(Y) relation was applied to estimate yield-dependent parameters for the von Seggern and Blandford RDP. The maximum-likelihood P- and PcPwave averaged spectra before pre-whitening are shown in Figures 7a and 7b. After correcting for instrument response, attenuation, and estimated RDP, the corresponding spectra are shown in Figures 7c and 7d. Because the P-wave was clipped at EKA for many of the events of interest, we performed the procedure on both P and PcP arrivals at the EKA array. The results are presented in Table 5, with uncertainties estimated from the scatter of the spectral levels

Teledyne Geotech

49

June 1988

20 SPECTRUM AVERAGE

CD

0 -

H c')

--

-

-

---

-D

-

-

-IJ

---

W

-

0.00

_____________ _

2.00

;

-

__

-____

Lt.00

_

6.00

FREQUENCY

_

_

_

_

8.00

_

_

_

10.00

Hz

P for NZ 1973270

Figure 7a. Maximum-likelihood P-wave averaged spectra for Novaya Zemlya event 73270 recorded at 19 EKA stations. The upper and lower error bars are indicated for each discrete spectral amplitude.

50

19 SPECTRUM AVERAGE

CDD

-oJ CO

-

0D-0

0

-0

(D

-U

D

- _~

~:b-

-

CDCCr-

-a

-

--

t

0.00

I

-

2.00

0~

1-

14.00

6.00

FREQUENCY

8.00

10.00

Hz

PcP Por NZ 1973270

Figure 7b. Maximum-likelihood PcP-wave average spectra for Novaya Zemlya event 73270 recorded at 20 EKA stations. The upper and lower error bars are indicated for each discrete spectral aplitude.

SI

0 I 973270 44-

LU

wL+

+ U

+f

+

+

+

+

Zb

+ -4-

+

+

+4+ +4

4 4$

-+

+-44-

+

+

++ +

+

+

I

.-

b.

'0 0.00

2.00

6.00

IfO00

FREQUENCY h V,--A VAt -\J

5.0 SEC

HZ

8.00

i0.00

/ VA 4

337 6 H NM O-P

Figure 7c. P-wave spectra for event in Fig. 7a recorded at EKA.

The spectra is

corrected for instrument response, attenuation, and estimated RDP. The horizontal bar indicates the bandwidth used to estimate the spectral amplitude of the phase.

52

.

. .

..

,.,.l . . ,.

... J.a,.

-

.i~

. ,..

J&,

..

. k

-...

.

.

.

0

197.3270

0

, _+

U)

4

+_. +

+

+

++

+++ ZL

1-4 -b

+

+

+

+4

+

+ 4

++ C-

+

4

44-

f

4

+

+++4

f

::: L)

+

+

+

;6+++

+

4--4

+

+

+

+

+

+ +

0++

+

II Z

--

I

II

0.00

2.00

Lt 00

FREQUENCY

5.0 SEC

I

6.00

8.00

10.00

HZ

f90 NM O-P

Figure 7d. PcP-wave spectra for event in Fig. 7a recorded at EKA. The spectra is corrected for instrument response, attenuation, and estimated RDP. The horizontal bar indicates the bandwidth used to estimate the spectral amplitude of the phase.

53

FINAL REPORT


TGAL-88-03

across the array of stations. The four southern events are denoted with an "(S)". The southern event identified as a double event by Lilwall and Marshall (1986) is denoted by an "(SD)" in the table. The average log(P/PcP) is estimated directly from the data for the four events with both P and PcP recordings on scale.

For all others events,

Log(P) was either extrapolated from Log(PcP) or the reverse. The arrows (=, and 4=) indicate the direction of extrapolation for each event. The value Log(P/PcP) = 0.69+0.16 was used for the southern test site and Log(P/PcP) = 0.91+0.09 was used for the northern test site. In the cases of the extrapolated Log(P), the stated error includes both the uncertainty in the estimate of the original Log(PcP) estimate and the uncertainty in the Log(P/PcP) estimate. The Log(P) estimates are plotted in Figure 8 against the Lilwall and Marshall (1986) mb for each event.

The largest events (1973255,

1973300, 1974306, and 1975291) may still be biased due to the censoring imposed by PcP-phase clipping at stations within the array and their consequent removal from the processing. For example, the southern event 1975291 clipped all but one station on the PcP signal and is therefore certainly biased low.

Teledyne Geotech

54

June 1988

I I

P and PcP from Novaya Zemlya recorded at EKA F

|r

,

,

.

!

-T•

-

,

V I-

3.5-

o3.0-

)2.0

P

I

-

1.5