UNCLASSIFIED

AD NUMBER

LIMITATION CHANGESTO:

FROM:

AUTHORITY

THIS PAGE IS UNCLASSIFIED

AD821664

Approved for public release; distribution isunlimited. Document partially illegible.

Distribution authorized to U.S. Gov't. agenciesand their contractors; Critical Technology; OCT1967. Other requests shall be referred to AirForce Technical Application Center, Washington,DC. Document partially illegible. This documentcontains export-controlled technical data.

usaf ltr, 28 feb 1972

■I

Is QO

FSO LONG-PERIOD L-ARRAY NOISE COHERENCE

20 October 1967

Prepartd For

AIR FORCE TECHNICAL APR'ICATIONS CENTER Washington, D. C.

•y

E. F. Chlburls

TELEDYNE, INC.

Under

199

Project VELA UNIFORM

Sponsored By

ADVANCED RESEARCH PROJECTS A6ENCY Nuclear Test Detection Office

ARPA Order No. 624

BEST AVAILABLE COPY

TFSO LONG-PERIOD L-ARRAY

NOISE COHERENCE

SEISMIC DATA LABORATORY REPORT NO. 199

AFTAC Project No.:

Project Title:

ARPA Order No.:

ARPA Program Code No.:

VELA T/6702

Seismic Data Laboratory

62U

5810

Name of Contractor:

Contract No..

Date of Cor. :ract:

Amount of Contract:

Contract Expiration Date

Project Manager

TELEDYNE, INC.

F 33657-67-C-1313

2 March 1967

$ 1,736,617

1 March 19 6 8

William C. Dean (703) 836-7644

P. 0. Box 334, Alexandria,Virginia

AVAILABILITY

This document is subject to special export controls and each

transmittal to foreign governments or foreign national may be

made only with prior approval of Chief, AFTAC.

This research was supported by the Advanced Research

Projects Agency, Nuclear Test Detection Office, under

Project VELA-UNIFORM and accomplished under the technical

direction of the Air Force Technical Applications Center

under Contract F 33657-67-C-1313.

Neither the Advanced Research Projects Agency nor the

Air Force Technical Applications Center will be responsi-

ble for information contained herein which may have been

supplied by other organizations or contractors, and this

documen is subject to later revision as may be necessary.

TABLE OF CONTENTS

ABSTRACT

INTRODUCTION

ORDINARY COHERENCE

RECORDED SIGNALS

CONCLUSIONS

REFERENCES

Page No.

1

1

Noise Sample #1 1

Noise Sample )K2 2

Noise Sample §3 3

MULTIPLE COHERENCE 3

4

5

6

ILLUSTRATIONS

Figure

1. Location Map of the TFSO L-array.

2. Ordinary Coherence Between TFO and PY1, PY2, PY3, PYU PY5. Noise Sample il.

3. Ordinary Coherence Between PY1 and PY2, PY3, PYH, PY5. Noise Sample §1.

4. Ordinary Coherence Between FY2 and PY3, PY4, PY5. Noise Sample #1.

5. Ordinary Coherence Between PY3 and PYU, PY5, and Between PYU and PY5. Noise Sample 11.

6. Power Spectra at the TFSO L-array for Noise Sample fl.

7. Ordinary Coherence Between TFO and PY1, PY2, PY3, PYU for Lag Windows of 200 Points. Noise Sample fl.

8. Ordinary Coherence Between PY1 and PY2, PY3, PYU, PY5 for Lag Windows of 200 Points. Noise Sample il.

9. Ordinary Coherence Between PY2 and PY3, PYU, PY5 for Lag Windows of 200 Points, roise Sample fl.

10. Ordinory Coherence Between TFO and PY1, PY2, PYU and Between PY1 and PY2 For Lag Windows of U00 Points. Noise Sample #1.

11. Ordinary Coherence Between TFO and PY1, PY2, PY3, PYU, PY5, With Tight Filter. Noise Sample il.

12. Ordinary Coherence Between PY1 and PY2, PY3, PYU, PY5, With Tight Filter. Noise Sample il.

13. Ordinary Coherence Between PY2 and PY3, PYU, PY5, With Tight Filter. Noise Sample il.

1U. Ordinary Coherence Between PY3 and PYU, PY5, and Between PYU and PY5 With Tight Filter. Noise Sample il.

ILLUSTRATIONS

(continued)

Figure

15. Ordinary Coherence Between TFO and PY1, PY2, Py4 PY5 Noise Sample #2.

16. Ordinary Coherence Between PY1 and PY2, PY4, PY5. Noise Sample #2.

17. Ordinary Coherence Between PY2 and PYU, PY5 and Between PY4 and PY5. Noise Sample #2.

18. Power Spectra At The TFSO L-array for Noise Sample #2.

19. Ordinary Coherence Between TFO and PY1. PY2, PY3 PYU Noise Sample »3. ...

20. Ordinary Coherence Between PY1 and PY2, PY3. PY4 PY5 Noise Sample )K3.

21. Ordinary Coherence Between PY2, and PY3, PY4, PY5 Noise Sample )K3.

22. Ordinary Coherence Between PY3 and PYU, PY5 and Between PY4 and PY5. Ncise Sample #3.

23. Multiple Coherence With TFO As Output and PY1 through PY5 as Inputs. Noise Sample II (upper) and Noise Sample f3 (lower).

21. Large Love-Wave Signals Recorded At The TFSO L-array.

ABSTRACT

Three long-period noise samples recorded at the TFSO

L-array were analyzed for coherence properties. The results

indicate that the ordinary coherence is generally high between

elements 5-10 km apart and low between elements further apart.

Multiple coherence is high for the first noise sample but low for

the third sample.

Zero-delay noise summations for an additional sample produce

about Ns improvement over the average RMS noise level and beam-

forming of a large P-wave signal produces about N^ improvement

in signal-to-noise ratio.

INTRODUCTION

During the period from about 01 February 19 67 to 06 April

1967, a six-element array of long period seismographs was oper-

ated in the vicinity of the Tonto Forest Seismological Observatory

in Arizona. The L-shaped array was composed of two legs bearing

ENE and SSE with lengths approximately 25 km and 15 km respectively

(Figure 1). Each of the sites contair-d three-component Geotech

Model 7505A vertical and 8/OOC horizontal seismometers (free periods

of 20 sec); photo-cell amplifiers were at all sites except TFSO

which has a standard photo-tube amplifier.

The purpose of the array was to record and analyze the spatial

properties (coherence) of long-peiiod noise in the vicinity of TFSO

with a view towards installing a permanent long-period kB km hexa- gonal array of seven elements.

ORDINARY COHERENCE

Noise samples from three different time periods were used for

computing ordinary coherence. These samples, designated "Noise

Sample #1" through "Noise Sample 13". are from the following time periods:

Noise Sample ^1 - 25 February 67 0801Z

Noise Sample §2 - 26 March 1967 1025Z

Noise Sample §3 - 05 April 1967 1330Z

Noise Sample |1. This sample contains U000 points digitized

(by Geotech) at two points per second and prefiltered with a band-

pass of 0.01 cps to 0.30 cps (half-power) with 18 db/octave rolloff.

The ordinary coherences vs. frequency between all pairs of seis-

mometers are shown in Figures 2 through 5. These coherences were

obtained using a lag window of 50 points. As shown on these figures,

the site PY-5 appears to have noise which is incoherent with the

other sites. It is believed that there were no instrumental dif-

ficulties with this site (data are normal, visually, and the auto-

spectrum agrees with the spectra from the other sites, as shown in

- 1 -

411 ■ J^-

o

o o

1.0

7

ÜJ

Z C

Ul

tr. <

TFO-PY I rro-PYt TTC-PY 3

Tf0-Py4 TfO-FYS

.06 JO .11 14 .16

FREQUENCY (cps) 16 .20 .« 24 .26

FIGURE 2. ORDINARY COHERENCt BETWEEN TFO AND PYl, PY2. PY3. PY4, PY5. NOISE SAMPLE #1.

PYI-PV3 pyi-py4

PYI-PY5

02 .04 .0« oa .ID .it .14 .ic FREQUENCY {cp)

.!• t0 22 .24 M

FIGURE 3. ORDINARY COHERENCE BETWEEN PY1 AND n u ^ r\

1.0

.• - m-py*

Tr

FREQUENCY (cps) FIGURE 4. OROINARY cnuPOrnrf rrrucrM DV9 AMR

JL1

ü z S ' UJ

i >- .»

PY3-PY4 PY3-PV8 Py4-PY8

I

.OS .10 It .14 .1«

FREQUENCY (cpt) t4

FI3URE 5. ORDINARY COHERENCE BETWEEN PY3 and PY4, DTTurru nvA «kin nvc jn T c r c «um r JII

Figure 6) so that tne noise at this site is perhaps due to local

winds, site emplacement properties, etc. Subsequent noise coher-

ences from different times will be shown which indicate similar

properties at PY-5 and to a lesser extent at PY-U and PY-3.

Additional coherences were :omputed from the same data, but

with lag windows of 200 points, for various (not all) pairs (Fig-

ures 7, 8, and 9) and of U00 points (Figure 10). There are no

significant changes in the computed coherences using the longer

lag lengths, except those anticipated, such as a general increase

in coherence with the longer lag windows (see, for example, Chiburis

and Dean, 1967a, b, c, d).

The characteristics of the bandpass filter applied to the

data used for the results shown in Figures 2 through 10 (0.01 cps

to 0.30 cps, 18 db/octave rolloff) were changed to 0.017 cps to

0.200 cps, 12 db/octave rolloff and coherences recomputed to note

any effects due to filtering. Figures 11 through Ik show the coher-

ence results of the tighter bandpass filter. The only significant

differences occur at the higher frequencies, as expectfd, with more

instability in computing the coherence.

Noise Sample #2. This sample contains 2910 points at one point

per second (digitized at five points per second and decimated) pre-

filtered from 0.010 cps to 1.00 cps (24 db/octave) and then low-pass

filtered with a high-cut of 0.L0 cps. Ten-percent lags were used

(290 points). The results for six pairs are shown in Figures 15

through 17. The noise between the array sites for this particular

sample appears highly incoherent, generally remaining below 0.50

at the lower frequencies. This result suggests that a simple array

summation should suppress the noise by a factor of N . A computer

program, LOPSAN, is available at SDL (R. A. Hartenberger, personal

communication) that measures the RMS noise level (and signal levels

ae well) of array data before and after summing, or beamforming.

Root-meün-square noise measurements from data recorded on 12 February

1967 at 0730Z, 0750Z, and 0810Z were made to determine the amount of

- 2 -

IO- I 1 ^y i

— m W4 PV 5 Tr»o

1 1

IO« /, A 1 V^\\\

-1 \\ rV^

h-M - \\

hik -A — I'J- 1

hmt 10*

M \

i /

/ N

\1 \1

—III ^ nil 1 « -asX v\

f if % 1 '-I» riji 1 ( w IS \l\ I'U ii

•i iMI N \

10»

II

00 OS 10 IS

FREQUENCY (cpt) to IB

FIGURE 6. POWER SPECTRA AT THE T^SO L-ARRAV FOR NOISE SAMPLE #1.

-:

0 00 .04 O«

FIGURE 7.

•0« .10 .It .14 .1« FREQUENCY (cpt)

.it tO tt .t4 tt

l!?'_:y3' ?r4 FOR LAG WINDOWS OF ?on POINTS. NOISE SAMPLE #1

100

Wl -PY2 ^ri-rrs ryi-^Y4 FYI-^Y»

Ot .04 04 OS tt J4

FIGURE 8.

10 it M w it

FREQUENCY (cpt) ORDINARY COHERENCE BETWEEN PY1 AND PY2 llhrl"-^* F0R LAG HIND0HS OF 200' POINTS. NOISE SAMPLE II.

PYf-rY4

it to tt 24

FIGURE 9

oa 10 it .14 .1« FREQUENCY (cps)

ORDINARY COHERENCE BETWEEN PY2 AND PV3. PV4, PY5 FOR LAG WINDOWS OF 200 POINTS NOISE SAMPLE II.

i.Or

tro-^ri T>o-rvt TFO -^4 PVI -m

04 04 14 I« SO It .t4 t4

FIGURE 10.

04 JO .11 .14 FREQUENCY (cpa)

ORDINARY COHERENCE BETWEEN TFO AND PY1. PY2, PY4, AND BETWEEN PY1 and PY2 FOR LAG WINDOWS OF 400 POINTS. NOISE SAMPLE #1

"öi .ö« .10 Tt I4 j« 1»

FREQUENCY (cpt) 20 22 24

FIGURE 11. ORDINARY COHERENCE BETWEEN TFO AND PY1. PYZ, PY3. PY4, PY5 WITH TIGHT FILTER. NOISE SAMPLE #1.

PYi-m

FrcPY4 PYI-PY8

Ot .04 .M .0« .10 It 14 .1« .11 FREQUENCY (cpt)

«0 tt t4

FIGURE 12. ORDINARY COHERENCE BETWEEN PY1 AND PY2, PY3, PY4, PY5 WITH TIGHT FILTER. NOISE SAMPLE #1.

e-

pvt-nr» PYt-r»4

oo. Ot .04 OS Oi .10 .a .14 16 IB FREQUENCY (cpt)

20 tt 24 26

FIGURE 13. ORDINARY COHERtNCE BETMEEN PY2 AND PY3, PV4, PY5 WITH TIGHT FILTER. NOISE SAMPLE II.

lü

i < z

i

pys-PV4 PVS-rYB PV4-rr8

.i -

.00 01

FIGURE 14.

.0« .0« .10 .It .14 .1« It JO .It 14 t« FREQUENCY (cpt)

ORDINARY COHERENCE BETWEEN PY3, AND PY4, PY5, AND BETWEEN PY4 AND PY5 WITH TIGHT FILTER. NOISE SAMPLE #1.

LO

TFO-nrl TFO-nrt TF0-PY4

TFO-FY»

.00 01 .04 .0«

FREQUENCY (cps)

FIGURE 15. ORDINARY COHERENCE BETWEEN TFO AND PY1 PY2. PY4, PY5. NOISE SAMPLE 12.

m-PY4

Ot 04 Ot 10 It .14 i« ia to Jt .t4 2« FREQUENCY (cps)

FI6URE 16. ORDINARY COHERENCE BETWEEN PYl AND PY2. Py4, PY5. NOISE SAMPLE #2.

10

PY2-PY4

PY4-PV6

ÜJ

z o g

FIGURE 17.

FREQUENCY (cpt)

OROIN/iPy COHERENCE BETWEEN PY2 AND PY4, PY5, AND BETWEEN PY4 AND PY5. NOISE SAMPLE «.

noise improvement possible with a zero-delay summation. This record-

ing date was selected on the basis of high apparent noise background.

Two combinations of array elements were used in the summations. The

following table gives these results.

Time Sample Elements

0730-0750 0730-0750 0750-0810 0750-0810 0810-0830 0810-0830

TF0,PY1,2,3,<*,S TFO.PYl,2,4,5 TFO,PYl,2,3,i+,5 TFO,PYl,2,i4,5

TFO,PYl,2ti»,5

iVg.RMS* Summation Improvement RMC db NJjdb

47.66 21.01 7.1 7.8 46.98 22.02 6.6 7.0 47.82 18.84 8.1 7.8 47.01 21.89 6.6 7.0 58.93 25.13 7.4 7.8 57.97 26.24 6.9 7.0

*Each data trace was "demagnified" by an arbitrary value to yield about the same individual RMS level.

The data recorded at PY3 appeared to be questionable so array

summations were made both with and without PY3 data. The improvement

results in either case are close to the predicted N^db indicat.tag that

the noise is spatially uncorrelated for zero-delay summations.

The spectra for Noise Sample #2 at f.-'ve of the array sites is shown in Figure 18.

Noise Sample #3. This sample contains 3574 points (digitized at

one point per second), band-passed from 0.017 cps to 0.500 cps (12 db/

octave), and low-passed with a high-cut at 0.35 cps. Again, 10% lags (360 points) were used.

22. The results for all station pairs are shown in Figures 19 through

MULTIPLE COHERENCE

Multiple coherences as a function of frequency were computed for

Noise Samples #1 and #3. Multiple coherence indicates the number of

input data channels which would be necessary to describe a noise field

and gives a quantitative measure, versus frequency, of how well a linear

combination of these n imput channels can match the (n + l)st channel.

The selected autpur channel for both noise samples was TFO and the in-

puts were PY1 through PY5. These results, shown in Figure 23, indicate

- 3 -

10 .It

FREQUENCY (eps)

FIGURE 18. POWER SPECTRA AT THE TFSO L-ARRAY FOR NOISE SAMPLE #2.

10

TfO- FVI .1 ^~ -^^ TPO TfO-

•rrt J • TFO- rr4

111 ____ TFO- FYB T

- 1

S . n > M

'

^ / A /\ ^/v // //

t \ m ̂ A/^A 4. /\ .1 i\/^Ä3 ^^^V AT ft M \J m ^^V y ii t \ / Xt \jiU J

r-r 1^ j^^fAJ^^A^^ ̂ /Vs & T 00 Of A4 0« M .10 II .14 it !• .tO .at .14 M

FREQUENCY (cp«)

FIGURE 19. ORDINARY COHERENCE BETWEEN TFO AND PV1. PY2. PV3. Py4. NOISE SAMPLE #3.

bJ

.00 Ot 04 .04 04 .10 .It .14 .1«

FREQUENCY (epo)

PM •rvt •ftt

— ^ _ _ pyt •rr4 i

•rrt

A ^

\

^ ̂ _ 14 JO It M4

FIGURE 20. ORDINARY COHERENCE BETWEEN PY1 AND PY2, PY3. PY4, PY5. NOISE SAMPLE #3.

1.0

FIGURE 21. ORDINARY COHERENCE BETWEEN PV2 AND PV3. PV4, «»YS. NOISE SAMPLE #3.

.00 .0* .04 .0« .0« .10 .It .14 .10 .11 .tO ft .t4 t«

FREQUENCY (cpt) FI6URI 22. ORDINARY COHERENCE BETWEEN PY3, AND PY4.

PY5 AND BETWEEN PY4 AND PY5. NOISE SAMPLE #3.

FREQUENCY (opt)

FIGURE 23,

0» .„ „ FREQUENCY (opt)

to t* M

that noise sample #3 is multiply incoherent (less than about 0.6)

at all frequencies, whereas noise sample #1 is highly coherent at

frequencies between 0.0 2 cps and 0.09 cps. The low multiple co-

herence for sample #3 shows that there are no (or few) linear filter

relations between the six elements in the array.

RECORDED SIGNALS

Figure 24 shows a strong Love wave recorded by the horizontal

instruments at the six sites in the array. Visually, each sice

appears to record very closely the same signal data on all instru-

ments .

A large teleseismic e/ent was recorded on 13 February 1967 at

about 2300Z. Program LOPSAN was used to determine the P-wave S/li

improvement from beamforming. Time delays were determined by eye.

The results are given below.

Element S/N

PY1 10.70

PY2 19.90

PYU 19.21

PYf. 26.79

TFO 20.50

Mean 19.42

Phased sum U?.72

db improvement 6.8

N^db 6.6

- >♦ -

c

C

■ J mm

It

i

CONCLUSIONS

1. Ordinary coherence within the passband is high (greater

than 0.8) between elements 5-10 km apart, low (less than about

0.6) between elements further apart. Noise Sample #2. however, has

low coherence between all sites at all frequencies.

2. The sites PY5 and PY4 are generally incoherent with the

other sites, probably due to local noise characteristics.

3. Multiple coherence is high (greater than 0.7) between

0.02 c.ps and 0.09 cps for Noise Sample *1 but low (less than 0.6)

at all frequencies for Noise Sample #3.

1. Z^ro-delay RMS noise summations produce about N*5 improve- ment over the average RMS noise level.

5. Beamforming produces about N^ improvement in signal-to- noise ratio.

- 5 -

REFERENCES

Chiburis, E. F. and W. C. Dean, 1967a, "Multiple Coherence of

Long Period Noise at LASA", Report No. 189, Seismic Data Lab-

oratory, Teledyne, Inc., Alexandria, Virginia.

1967b, "Multiple Coherence of Short Period Noise

at LASA", Report No. 190, Seismic Data Laboratory, Teledyne,

Inc., Alexandria, Virginia.

1967c, "Multiple Coherence at 3 Vertical Arrays,

UBSO, GV-TX, AP-OK", Report No. 191, Seismic Data Laboratory,

Teledyne, Inc., Alexandria, Virginia.

1967d, "Multiple Coherence of Short Period Noise at

UBSO and TFSO", Report No. 192, Seismic Data Laboratory, Teledyne,

Inc., Alexandria, Virginia.

- 6 -

Unclassifled

DOCUMENT CONTROL DATA • RID •M ••*«<« WfM« M «iMaKl*«

TELEDYNE, INC. ALEXANDRIA, VIRGINIA

Unclassified Ik ««Ou»

I MPMT TITLI

TFSO LONG-PERIOD L-ARRAY NOISE COHERENCE

cientific t AUTMttnn) OM« MM. ISl

Chiburis, E. F,

«.MMMUVI October 20, 1967

TT^S 34 H

• a CaMTMACT O» •■*«? M«.

F 33657-67-C-1313

VELA T/6702

ARPA Order No

HUtf*t*>

199

«ARPA B2H

Program Code No. 5810

4,M«""I,J"T "OW Mhr 1*1»«

ogra TT97CM I* * v* iL AaiLiTY/LiMiUri^N *6heäi

This document is subject to special export controls and each trans- mittal to foreign governments or foreign national may be made only with prior approval of Chief, AFTAC.

II. »UP«. tmHTAHY MTU it. t»OMom«« mux Any AC riviT» ADVANCED RESEARCH PROJECTS AGENCY NUCLEAR TEST DETECTION OFFICE WASHINGTON, D. C.

It AMTMACT

Three long-period noise samples recorded at the TFSO L-array

were analyzed for coherence properties. The results indicate that

the ordinary coherence is generally high between elements 5-10 km

apart and low between elements further apart. Multiple coherence

is high fjr the first noise sample but low for the third sample.

Zero-delay noise summations for an additiotal sample produce

about N* improvement over the average RMS noise level and beam-

forming of a large P-wave signal produces about N* improvement

in bignal-to-noise ratio.

DD HE 1473 iManttijtf—- SäcäritvCUnia C«BOB

Unclassified Secmity CI«»»ific«tion

«ft «OIIOS LINK A UNK B ■ OH wt

UNK C HOL« WT

Noise Coherence

Seismic Arrays

Tonto Forest reismological Observatory

iramucTiom I. ORIGINATINO ACTIVITT! KM« th. MM «IH «Mm* •f <lw conlractM, •dbconlraclor, (rani**, DvpartaMnl of D»- <»—» «rll»!!» oc SIlMr orgsnliallon (coiporf milhmt) l»uln(

la. REPORT SFCUHTYCLAWIFICATION: Bain Ik* •*•*> •II Mcarlly claaalftcallaa of dia rapaM. IndtcMo nlmlm "RaaMclad Data" la laclada* MaAlag 1« to bo lo accar* aaca wilh apuapHal« oocurll* ragalulaaa.

2». GROUP: A*nmmic downcr^dlnc la «poclflod la DeD Dl- racllva MOO. 10 tm4 Arm** Patcaa Indtiolilol Moouol. I mar Ika coup a«a*«t. Aha, «kaa applicabla. ahow Ikal optional aMrklnga hay* baaa uaod (or Otoup J and Oreap 4 as aulhor- iü* J. REPORT TITLE: Srtot Ika caa»lala raport Ittto la all capital iMtata. TlUaa la all eaaaa alwald ba onclaaaUlad. ■ a awaalncTal tula cannol ba aalactad wlthoal daaaiflc» tlon. ska« tlila daaairiealtaa la ail capital* la panalkaala latiidlalaly Mlawlag Ika Ulla.

4. DOCmPTlVB NOTBft If raparl, ».»., IMaHai, progr***, Gl>« Ike ladaalva dal** «baa a

, «nt«. ika lip« of •al, w llaal.

Ic rapattlag patled la

S. AVTHORft)! Em*f Ik« a*aM<a) at aulkoK*) aa ahown aa at la Ika raport. Balai laM naaw. flral aaaw, aiiddla lalllaL It aulHwy, «ka« rank and bcanck «f **rvlc*. Tba naaw at Ik* principal >Mhor ia aa «baelata alalaMMi ra^ulraawnt

*. REPORT DATE; lalar Ika d*l* of Ik« roport a« da». noniK r»*r. at «walk. yaar. If anta than aa« dal« *pp**r* «a Ik« rapoct, «aa data of publlcMton.

7a. TOTAL NUHBER Of PAOBft Th* total p*«* eaaal •kmid folio» aawail paalaMlaa pro^odw**. La., aalar lha numbaf «f pac** caalalalai lallfallia

Tb. KUMBER OP REPERBNCEft Balar th« lalal aaa*«r of r«r*r*tK** cH«4 la Ik« raport.

•a. COIfTRACT OR GRAKT HUMBBR: It «proprlal*. aal« Ika appllcaMa aaatkar at Ik« caattac« or «raat uadar «kick 11M Kpoft WAS WlttMh

U. tc, h far PROJECT Wf Balar II nllllary d*partaaal Hwlllkallaa, lack a« pretad ■atprejact au*di*r. «jraUai aaakara. l*ak aaaMr, ale.

ta. ORIGINATOR'S REPORT NUteEIKS): Balar Ika afB- clal rrpert awabar kf «kick Ika «iwial »Ul ka ld*«lfl*d aad conlrallad br Ik« erlclaalla« activity. Thl* aaakar M*! k* aat«H I« Ikta raport.

M. OTHER REPORT HUMBBIKf): It Ik« raport ka« ka<a ••al(aad an» olhor Mport niaabar* (tllhmr by lit» orfd/n*«or •r be tk« apomorl, all« aatar IM« aaaAaKa).

ia AVAILABILITY/LIIHTAT10N NOTICE* Bataraaylk» Halloo* aa furllwr dl***ula*llon «I Ik« «port, «tkar Una Ik«««

kapaaad by ««curlly claaaitlc* ■aek *•:

landard atataaMnt*

(1) "QuallfUd r*q«*M*r« Bay aktala coplaa of tkta .-•part fro«, DDC."

(2) "Foralfn anKeuncaannl and dlaaaailaatiea of Ihla trpori by DDC I* not anlhorttaA"

(3) "U M. Gavanuaaid agancl*« may «Mala copla« of •hi« roport dlracüy fro» DDC Otbar «aal Iliad DOC •aara a hall rrijuaai tbraagk

(4) "U. S military *fla«cl*a nay obtain copla* of thl* raport dlractly from DDC Other qaallflad aaar« ahall raqural lhi*«gk

(S) "All dlaMbutlo« at Ibis raport Is coatrallad. Qtt«l- lfl*d DDC «Mrs ahall raquaat throu(h

If Ik« roport ha* kaaa Imlahod Ic lha Ottlca of Tactadcal Satvlc*«, Dapartawnl at Comarca, far aala to lha pidtllc, Indl- cal« IM* l*ct aad aalsr Ika prlr*. If known,

IL SUPPLEMERTARY NOTES: Us« for additional «aplsaa- Isry notoa.

11 SPONSORINO MIUTARY ACTIVITY: Erder Iks aaaw at Ihn d^aitSMBlal projacl of lie* or laboratory •ponaorlnf Cpar laf t) Ika r****rch aad datralopaiaat Inclad* addr-a*

l*. ABSTRACT; Balar aa abaliacl fl»lnf a bflol aad f*ctu*l sataaary at lha dacutaaal Indlcatlva of Ika report, araa though It stay alaa appaar al**wk*i« la Ika bod« of Ika lackalcal ra-

il addition*! spaca la raqulrad. a c allausllaa ah**l «kail

U I« klahly d**ir*bl* that Ik« abatrari »f claaalflad raporta k« uoelaaalflad. Each parapr ph of Ik* abalnicl ahall aad with aa ladlcatlon of Ika military sacarlly cUsslflcstiaa of Ihr la- fonaallon la Ika paragraph. Mpra**nl*d aa (TU, (th <C). ar tu>

Tkara I« a« Hallatlaa en Ik« length of th« akaüact. How- •Mr, th* satssslsd Imgth la tnm 150 to 22S trad«.

14. BBT RORDS: Bay word* gr* t*chnlc*lly SMsalacftil tana* ar «hort pkiaaas Iksl charactarUa a »port «ad aisy ka «aad as lad«! «aMaa for cataloging Ola raport Bay word* ara*t ka salscMd «« «Ml a« aacurlty claaslflcsllaa Is w^-lrad. '

«ack a« aaulpnaat taod«! d**lviatloa, Irada astas, ■llltaty 4 cad« aaais, fsagnpklc location. My ba i»a.| as bay

»ord* but will ka tellaatad by aa Indication of t.chnlcal con- last. Th* aaalcaarrat at link*, nil**, «ad w*lghta Is opUonal

Security Clt rily ClaMiflcatlM