607:

XFWS-A607 1-26(1970)

U.S. Fish VVildl, Serv.

Spec, Sci. Rep. Fish.

Studies on Continuous TransmissionFrequency Modulated Sonar

SPECIAL SCIENTIFIC REPORT-FISHERIES Na 607

SPECIAL SCIENTIFIC REPORT-FISHERIESEDITORIAL STAFF

Leslie W. Scattergood, Editor

Mary S. Fukuyama, Associate Editor

PUBLICATION BOARDJohn A. Guinan

John I. Hodges Edward A. Schaefers

Harvey Hutchings Parker S. Trefethen

John M. Patton, Jr, Robert C. Wilson

Leslie W. Scattergood, Chairman

Special Scientific Report— Fisheries are preliminary or progress reports

and reports on scientific investigations or restricted scope. Established as

Special Scientific Report in 1940, nos. 1 to 67 were issued from that date to

1949, when the new series. Special Scientific Report— Fisheries, with new serial

numbering, was started.

Special Scientific Report— Fisheries are distributed free to libraries, re-

search institutions. State agencies, and scientists.

UNITED STATES DEPARTMENT OF THE INTERIORWalter J. Hickel, Secretary

Leslie L. Glasgow, Assistant Secretary

for Fish and Wildlife, Parks, and Marine Resources

Charles H. Meacham, Commissioner, U.S. FISH AND WILDLIFE SERVICE

Philip M. Roedel, Director, Bureau of Commercial Fisheries

Studies on Continuous Transmission

Frequency Modulated Sonar

FRANK J. HESTER, Editor

United States Fish and Wildlife Service

Special Scientific Report--Fisheries No. 607

Washington, D.C.June 1970

CONTENTS

Page

Sonar Target Classification Experiments with a Continuous -Transmission DopplerSonar, by Frank J, Hester 1

Introduction 1

Fish echoes in relation to sonar detection 1

Operating principles of a CTFM sonar for fishery work 2Controlling relationships 3Doppler system in block diagrann 3Target strengths of single fish 4School targets and their acoustic properties 5

Feasibility study 6Lake Murray equipment tests 6Swimming pool tests 7Evaluation at Sea World 8Mission Bay Yacht Club pier 8Catalina tests 9

Feasibility tests of methods of data evaluation and reduction 10The DELTIC spectrum analyzer HKay vibralyzer and DELTIC 11U.S. Naval Ordnance Test Station 12Slow scan analysis technique 123-D technique 13

Conclusions from feasibility study 14Shipboard target classification experiments 14

Equipment 14Methods 15Results of shipboard tests 17

Reconnnnendations 18Sumn-iary 20Acknowledgnaents 20Literature cited 20

Acoustic Target Strength of Several Species of Fish, by H. W. Volberg 21

Introduction 21Facilities, equipment and methods 22Tests 23

Fish without swim bladders 23Fish with swim bladders 24

Relation of present to earlier findings 25Summary 26Acknowledgments 26Literature cited 26

Sonar Target Classification Experiments with a

Continuous-Transmission Doppler Sonar

By

FRANK J. HESTER, Fishery Biologist

Bureau of Commercial Fisheries Fishery-Oceanography CenterLa Jolla, Calif. 92037

ABSTRACT

A continuous -trans mission sonar with very fine echo frequency discriminationwas designed and constructed to study Doppler effects caused by the motion of fish

as they relate to fish size and swimming characteristics. Although the equipmentperformed as theory predicted, difficulties with sea noise and trouble in maintainingcontact with fish schools showed that commercial application of this approach is

unsuitable without considerable additional development work. These problems andsome results are discussed, and notes on target-strength measurements of severalspecies of fishes are included in this report.

INTRODUCTION

In 1961 the Inter-American Tropical TunaCommission recognized that the stock of

yellowfin tuna in the eastern tropical Pacificwas in danger of being overfished and pro-posed a catch quota to conserve the resource.This quota could be increased if fishing for

small yellowfin tuna (fish less than 20 to 25pounds [10 kg.]) could be controlled (Inter-

American Tropical Tuna Commission, I960).

It is difficult however to determine the size of

the fish before it is caught, and the attempthas never been made. In addition, it is of valueto the fishermen to know approximately the

size of the fish before setting their nets. Thisinformation decreases the chance of catchingfish so small that they "gill" in the net meshes.

Recognizing the economic and managenaentbenefits to be gained fronn determining the size

of a fish before it is caught, the Tuna Re-sources Laboratory of the Bureau of Com-mercial Fisheries began a study to develop a

shipboard sonar that could locate and classifyfish schools. This study began in 1963 andended in 1968. The work was done chieflythrough contracts with industry. Straza In-

dustries,^ now Ametek/Straza, was the primecontractor, and the bulk of this report is takenfrom their statements of progress. For thatreason Frank Hester has preferred to belisted as Editor.

^Use of a trade name does not Imply endorsement bythe Bureau of Commercial Fisheries.

FISH ECHOES IN RELATION TO SONARDETECTION

One of the first items of concern was whatgeneral categories could be used to classifyfish. Table 1 gives the general categories forthe various means that might be used to detectand identify various single fish specimens.This table does not represent all of the possiblemeans but does include the obvious techniquesthat might be used for the detection and identi-fication of fish. The visual observations columnof the table includes movement in two cate-gories: velocity and body motion, i.e. withrespect to the mediunn and with respect to thebody of the fish. The acoustic observationscolumn is included only to show that manycategories of noise may be produced in the seafrom animal life. These sounds can interferewith acoustic observations, even with ultra-sonic equipment. In the frequency range 50to 60 kHz (kilo Hertz); this problem is notserious but sea organisms can producecertain noises that interfere with measure-ments .

Table 1 indicates three separate categoriesunder the active acoustic heading. The first ofthese categories, the static echo return ortarget strength of a fish, were measured on alimited number of specimens at the U.S.NTCF (Navy Transducer Calibration Facility)at Sweetwater Dam, San Diego, Calif., to de-termine the target strength and directivitypatterns from various specimens. This work,which was necessary for specifying the design

Table 1. --Methods of identifying single fish by visual and acoustic observations.

Visual observations

P*»OJECTO«

T

sonar system, if the sound is reflected from amoving object and the receiver is stationary.In table 1, column 4, items c, d, and e deter-mine the character of a Doppler signalreturned from a fish moving with respect to

the medium. These parameters are concernedwith the fish's position and velocity componentswith respect to the hydrophone and projector.Superimposed on the total or average velocityof the body of the fish would be the minute Dop-pler changes produced by body flexure. Thesebody flexures are necessary for propulsion of

the fish or are a result of the propulsion orrespiratory mechanisms. Item a in column 5 ofthe table lists this characteristic. The relationof tail beat to velocity is one of the factors thatinfluence Doppler effects.

Figure 3 shows the block diagrann for theDoppler system constructed as a part of thisprogrann. Many of the parameters for thisDoppler system were adjusted to be compatiblewith the hydrophone projector and receiver ofthe fishery sonar. The sensitivity of such asystem is the sanne as the sensitivity of thesonar itself. One difference is that range de-termination can no longer be made. Since thesystem is only a velocity-measuring device it

is valuable only in determining whether or notmotion with sufficient target strength is gen-erated. Because only a portion of the fishproduces Doppler effects due to the swimming

motion, the information content is weaker thanthat in a normal sonar return. The theoreticalvoltages present in the Doppler receiver arenoted in figure 3.

Target Strengths of Single Fish

From the work done at the NTCF, Strazadetermined that the target strength on a beamaspect for individual fish was about -25 db, onthe average. Some of the specimens tested in-dicated target strengths as low as -34 db. Sincethe target strength for a fish without a gasbladder apparently is a function of the projectedarea of the fish, it is important that the fishbe of reasonable size for the evaluation of thesignal returns. In all the experiments the re-turns from the fish were large enough for thesonar to detect and give information on its

range. The Doppler unit is concerned onlywith that fraction of the fish that is moving at

the velocity that can best identify th^ motion ofthe species. The signal level for the Dopplerunit is therefore less than the signal for thesonar. Enough infornnation was gained in thesetests to indicate that with larger fish theDoppler signal would be great enough to permitacquisition of Doppler data. Furthermore, anincreased number of fishes, such as a school,would increase the signal-to-noise ratio of the

system and pernnit rapid evaluationof the data.

School Targets and Their AcousticProperties

The theory applied here with regard to fishschools is very similar to the general approachused for measuring the scattering and absorp-tion of sound in wakes. No dependence is placedon gas bladders within the fish, however, andthe target characteristics for the individualfish targets are those obtained by both dynamicand static tests at the NTCF. For lack of

additional information regarding the acousticproperties of the fish, the effective absorptioncross section, ^ o , is assunned to be aboutequal to the scattering across section, a «

obtained from these tests.

The assumption is made that the fish schoolis dispersed so that the distance betweenneighboring fish is significant. The wholeschool contains many fish, and some of thesefish will not be in the sound beam. Only thosefish in the intersection of the sound beam andthe school contribute to the acoustic effects.

The total number of fish that are effective in

producing an echo depends on the width of the

school, the angle between the school andthe axis of the sound beann, and the

area of the sound beam that the school in-

tercepts.The calculations involving wide schools are

different from those which involve narrowschools. In a sonic view of a narrow school,the relative total nunnber of fish in the active

reflective volume is small, so that the pro-jected areas of individual fish overlap onlyrarely. When this condition is fulfilled the

school is considered to be narrow. On the otherhand, when there are so many fish that theirprojected areas usually overlap, the school is

considered to be wide.To obtain a useful approximation for the

school target strength the beam pattern of the

sonar set nnust be included in the calculation;involved is the area. A, of the sound beamintercepted by the school. The accurate calcu-lation of A is complicated, and the values of

many of the quantities entering into it areuncertain. Therefore, a very rough calculationwill suffice for the present discussion. If theangular half-width of the sound beam is

J3

radians in a given plane, then the width of the

beam at a range of R yards from the sonarwill be 2 R yards. ^ If the school has a verti-cal dimension of Aq yards which is less thanthe vertical width of the beam at the school,an approximate expression for A is

2R0Ad

Note: Because the contractors Involved In this study

habitually use the English system of measurement In

sonar work this notation will be used throughout. Themetric equivalent will be given In parentheses whenappropriate.

The above expression may be substituted into

the general target strength expression for the

school, which is given as

N a AwTs = 10 log( 4t cos e\

where N = average number of fish in a unit

volume of the school (l/yards ),

(T = scattering cross section o£one fish

(yards^) constant for angled ,

A= 2 R )? A Q (yards^),w= geometric width of the school

(yards),

d = angle between axis of sound beamand perpendicular to the axis of

the school.

Introduction of the area expression into theabove equation and separation of the expressioninto two terms give the resulting approxinnatetarget strength for the narrow school

Ts= 10 log (^1^:^). 10 log (i^).TF cos 6

The first expression contains only the quanti-ties characteristic of the school whereas the

second involves quantities describing the posi-tion of the sonar and the bearing and width of

the sonar beam. The first term in the expres-sion can be referred to as the strength of the

school and nnay be interpreted as the targetstrength of the school for 1 yard (0,9 m.) of

the school.In wide schools, the total number of fish in

the active volume is so great that overlappingof the projected areas is extensive. If the

foregoing equations were used to calculate the

power removed from the sound beam, the endresult would be that the school removed morepower than was incident upon it. This obviousimpossibility results from the neglect of the

overlapping projected areas in the previousequations.

The overlapping can be accounted fpr byassxaming that the fish nearest the source castsshadows on those farther away, removingpower from the incident sound beam or in the

returning echoes, A rough approximation forthe foregoing considerations results in anequation for the target strength of the wideschool.

T = 10 logAa

8t<t,

1 - e

-(2N(T w/cos9)

The foregoing expression is still somewhatrough in that smaller fish would not cast sharpshadows (depending upon the wavelength).Furthermore, second-degree scattering fromone fish to another is ignored and only soundscattered once is considered.

The above equation can be used with the

following assumptions: R= 1,600 yards.

a = 0.04 yard , 6°, 2^ = 0.1 radian horizontalby 0.2 radian vertical, A d = 20 yards and N =

l/yard-'. By use of the foregoing, the targetstrength for the school appears to be about 19

db. This approach also considers the fact that

the vertical dimension of the beam is widerthan the depth of the school. Aq, thus the

vertical dimension is 20 yards. Also, an in-

cremental width, Sw, of 16 yards is taken to

agree with a typical CTFM sonar resolution.

Checking this result with the summation of fish

targets as determined from the target-strengthwork at NTCF we obtain a target strength in

the order of 16 db. From this rough checkand the above solution to the equation atypicalschool target strength of about 15 db appearsto be a very conservative value, for the para-meters selected.

To examine what the foregoing schooltargetstrength represents in terms of a CTFM sonarset of the same beam pattern as selected in

the foregoing expressions, we can use the sonarequations as follows:

2Nw

-30fl (9 1m.)-

BUOY

Figure 4.—Test arrangements at Lake Murray.

find the best method for removing weak targetinformation from the recorded data. Althoughwater tests had been made in the very smalltank at Straza Industries, it was not possibleto project sufficient radiated power into thewater for proper evaluation of the operatingperformance of the equipment. A tank wouldnot have been suitable for this series of testsbecause reverberations would influence thereceiver performance. Lake Murray, a reser-voir within 7 miles (11,3 km.) of Straza In-dustries, was made available for a test throughthe cooperation of the San Diego County Fishand Game Service. This enclosed body of waterwas free of manmade noise on 2 days of theweek while closed to the public. Lake Murrayseemed to meet the evaluation requirements.The equipment was transported to the site, andrecordings were made of various inanimateobjects.

To evaluate the performance of the equipmentappropriately at Lake Murray required a targetof a known signal strength. Such a target, atriplane 4 inches (10 cm.) on the side, wasbuilt. Equipment was built to support thetriplane and also to permit it to rotate in a4-foot (1.2 m.) circle (fig. 4).

The first returns from the snnall targetsused in the Lake Murray tests indicated thatthe receiver was not performing with therequired sensitivity. The receiver was there-fore redesigned on the basis of the informationgained in these tests. Actually, the receiverperformed well but not to the required speci-fications. The receiver did respond marginallyto small fish swimming near the transducers.

A second test at Lake Murray indicatedthat the receiver was performing with thedesired sensitivity and could respond to ex-tremely small targets. This finding, however,did not necessarily mean that the sonarequipment would perform well on fish in smallcontained bodies of water. After the secondtest the next step was use of the equipmenton live fish of the desired dimensions andspecies that were to be studied in the program.

Swimming Pool Tests

Before tests were started with live fish in

a large body of water, a series of tests weremade to evaluate various enclosed sites. Onetest was carried out in a kidney-shapedswinnming pool to evaluate the crosstalk orreverberation produced by the walls of the

pool. The level of crosstalk was found literally

to overwhelm the receiver. The reverberationlevel was so high that it nearly blotted out

large signal returns produced by an 18 -inch(46 cm.^ square flat reflector of aluminiim.It was impossible to detect the small triplanesthat had been used in the Lake Murray tests.This first indication that the pool wouldproduce substantial reflections from its wallswas not totally unexpected, but the level of thereverberation was surprisingly high. Beforea test site could be decided upon, it clearlywas necessary to transport the equipment to

the site and determine the reverberation level

of the tank. Since this test had been in a

kidney-shaped pool it was possible that a

pool with a different shape might produce

reverberation levels which did not prevent thedetection of the fish. The best possible shapefor the operation of the sonar would be a tanksimilar to the USNEL. Transdec transducertest facility but this tank was not available.Therefore, the next possibility was a tankwhich would scatter the sound and preventdirect reflection from the walls into the hydro-phone and sonar. The only shape that wasavailable and could possibly perform thisfunction was that of the circular holding tanksat Sea World, an oceanarium located onMission Bay in San Diego.

Evaluation at Sea World

Permission was granted by M. Shedd, the SeaWorld Director, for an evaluation test to bemade in one of the circular concrete holdingtanks, 40 feet (12.2 m.) in diameter and 10 feet(3.0 m.) deep. We checked the holding tankfor reverberation at all angles of incidenceand depth of the transducer-projector array.The reverberation level was found to beexceedingly high and of such a character as tomask small targets conopletely. Using thetank for the evaluation of signal returns fromfish was therefore impossible. This problem,even though learned early in the program,raised considerable difficulties because nowit forced operation to an open-water environ-ment. The problem of containing a fish at somepoint in the sonar beam in open-water testsis more difficult.

Mission Bay Yacht Club Pier

For open salt-water experiments a site wasselected in Mission Bay, a shallow bay northof San Diego. A series of tests was made at

the Mission Bay Yacht Club pier to evaluatethe signal return from various materials thatcould be used for containment of fish specimensin open-water tests. Mission Bay was selectedbecause during a normal week day the boatactivity within a radius of 1 mile (1.6 km.) is

very low. Tests could be suspended during theinfrequent and brief passage of motor launches.The tests were made to determine the reflec-tion capability or target strength of severalmaterials. The materials tested included:4-inch-mesh (10 cm.) nylon net; 1-3/8-inch-mesh (3.5 cm.) nylon net; and sheets ofpolyvinyl, 0.002-inch (0.05 mm.) and 0.007-inch (0.18 mm.) thick. In all tests, the netmaterial gave signal reflections high enoughto mask small targets viewed through thenet. The sheets of polyvinyl material passedthe test with only small reflected signals.Most of the reflections were due to clingingair bubbles, which give extremely strongsonar returns. A 4-foot (1.2 m.) diametercylindrical, polyvinyl bag 15 feet (4,6 m.)long and suitable for operation in the wateradjacent to the floating pier at the Mission

Bay Yacht Club was constructed. This bagwas perforated with several thousand 3/8 -inch(9.5 mm.) diameter holes to allow oxygenatedwater to enter the bag during tests. A boomwas constructed to hold the polyvinyl bag atthe necessary distance from the pier. Figure 5shows the top view of the test site and indicatesthe location and distances in the test on severalfish specimens. The problem of obtaining testspecimens was very difficult because thewaters off and near San Diego had a red tidethat repelled fish. Therefore, the polyvinylenclosure was evaluated in a series of testswith fish that were not necessarily those ofgreatest interest in this program. The fishthat were used did give sufficient informationfor the equipment and data-processing tech-niques to be further developed.

During the first of these tests aSCUBAdiverobserved the position of two fish--kelp bass(Paralabrax clathratus )- -within the enclosure.The procedure was as follows: a specimenwas placed in the enclosure; the enclosure wasmoved into the sonar beam by rotation of theboom at its pivot point; the diver entered thewater, observed the fish's position, attitude,and behavior; the diver left the water. A datarun was taken; the diver reentered the waterand observed the location of the fish; it wasassumed that the behavior was reasonablyconsistent during the two observations. Theactual location of the fish within the enclosureduring a test was never determined owing tothe poor visibility under the red tide condition.The fish tended to go directly to the bottom ofthe bag; when they did they were no longer inthe major beam of the sonar. In subsequenttests a 6-foot-diameter (1,8 m.) bag wasconstructed with a "false" bottom heat-sealedinto the walls of the polyvinyl cylinder. Thisarrangement linnited the range of the fish in thebag to about 8 feet (2.4 m.) and ensured re-tention of the fish in the major lobe of thesonar beam.

The results of the fish tests with the diverindicated a problem in reverberation, not withthe walls of the enclosure as had been deter-mined in the previous tests, but between thesurface and bottom reflections in the shallowwater, which was 15 feet (4.6 m.) deep at

high tide. This problem manifests itself in theoutput of the equipment as multiple returnsfrom a single target. The tape output forexample showed two strong returns from oneSCUBA diver. Rather than nnove the operationto deep water, we attempted to determine thesignature of fish within the enclosure in thepresence of the surface and bottom reflections.Unfortunately, the returns were confused tosuch a degree that suitable presentation of themotion of the fish was impossible. The taperecordings carried indications of fish motion,but certainty as to the behavior of the specimenin the sonar was not possible. Further tests ofthe CTFM sonar were abandoned and work was

SONARHYDROPHON

aPROJECTOR

SONARCABLES ^-,-

OELTICr?:

ANALYZER EDTsu

SONARRECORDINGEQUIPMENT

POWERPPLY

UOv.o.c,POWERLINE \

POLYVINYLBAGENCLOSURE

DIRECTION OFROTATION FORLOADING SPECIMEN

Figure 5.— Top view of test site Mission Bay Yacht Club pier.

limited to the technique of presenting informa-tion on the Doppler effect of the fish, a processby which the signal return could be removedin the presence of the surface and bottomreflections.

Preliminary tests in air and pools of theDoppler portion of the sonar equipment hadbeen made at the Straza sonar laboratory.The operation was fo\ind to be extremely good;the low-frequency presentation which waspossible with the Doppler technique went to at

least 2 Hz, and was capable of operation in the

lOth's of Hz of Doppler information. A live,

76-cm. yellowtail ( Seriola dorsalis ) was pur-chased from one of the commercial fishing

boats and brought to the Mission Bay YachtClub pier. This fish was placed in the new6-foot-diameter (1.8 m.) bag. During this

test, the Doppler unit was used to determinethe feasibility of operation in the shallowwater. The tape recordings made at this test

were observed visually on a Tektronix oscil-loscope and aurally. The information that a

fish was in the enclosure showed up asdifferences in the appearance of the waveformsand changes in pitch of the aural signals.

Sophisticated techniques for data reductionwere required to extract from the taped infor-mation the signals from the fish and thosefrom the surface motion of the water. Thispoint is covered in a separate section--DataEvaluation and Reduction. The tests at the

Mission Bay Yacht Club pier were concludedand preparations for a deep, open-water test

were begun. Deep water was required becauseripples on the surface of the water and thereflection from surface and bottom caused toogreat problems. The intent was to make a test

in water deeper than 50 feet (15.2 m.) andthus eliminate the problem of surface andbottom return and at the same time permitthe operation of both the CTFM sonar and the

Doppler equipment,

Catalina Tests

A test site at Catalina Island was selectedfor several reasons: (1) clarity of the waterthat would permit films to be made of the

fish during the test; (2) availability of a large

number of fish in the size range of interest in

the area; (3) depth of the water sufficiently

great to eliminate most of the problems fromsurface and bottom reflections.

Both the CTFM and Doppler sonar equip-ment were taken to Catalina. These tests hadmany operational difficulties owing to the

operation of breadboard equipment on the

45 -foot (13.7 m.) charter boat Duchess whichwas used at Catalina Island (and engineer'smal de mer). Three SCUBA divers set up the

equipment on the bottom. The arrangement of

the equipment and the checkout and operationwent smoothly--all systems checked perfectlyand on the second day the television equipmentoffered by the U,S. Naval Ordnance TestStation, Pasadena, and all the sonar equipmentwere functioning perfectly. During the secondday the divers obtained abalone and placedthem on the chumming platforms so that fishwould accumulate in the beam of the sonar asdepicted in figure 6. The sonar returns wererecorded on one track of a Magnecord taperecorder. Data runs were taken for 1-1/2days with both the CTFM sonar and the

Doppler sonar equipment. A series of 14 film

MRM-503 MARKER RECEIVER

50ft (15.2 m.)

Figure 6.--Test arrangements at Catalina Island.

takes of 10 to 15 seconds duration were madethrough the use of a diver-operated l6-mm.Bell and Howell camera. Attempts at photo-graphing the television monitor screen wereunsuccessful owing to lack of synchronizationbetween the frame rate of the monitor. Toprovide time synchronization between the taperecordings and the motion picture film, a

crude but effective technique was employed:a Straza Industries MRM-503 Marker Re-ceiver was used for listening to the noise thecameraman made when he tapped a screwdriveragainst the face plate of the camera. Themotion picture shows clearly the screwdriverstriking the face plate of the camera housing,and the resulting click was recorded on thesecond track of the tape recorder. The distancefrom the diver cameraman to the hydrophonewas 60 feet (18.3 m.). The time lag betweenthe arrival of the pulse and the occurrenceof the event was therefore constant and lessthan one camera frame (24 frames/second).The conclusions from the Catalina tests aregiven below. After the processed l6-mm. filmfrom the Catalina tests was viewed, the variousfish species of interest were identified asCalifornia sheephead (Pimelometopon pul-chrum ), kelp bass (Paralabrax clathratus).

senorita (Oxyjulis californica ), and halfmoon(Medialuna californiensis).

FEASIBILITY TESTS OF METHODS OFDATA EVALUATION AND REDUCTION

The following paragraphs delineate the var-ious techniques that were used to extract thedata from the tape-recorded tests from MissionBay and Catalina Island. The associated blockdiagrams and setups of each method areincluded. Several techniques were used in aneffort to find the best way of removing theinformation. In all of the presentations exceptone, that of the DELTIC (Delay Line TimeCompression), the time that was necessaryfor the evaluation is indicated. The reasonfor this exception is that DELTIC operatesin real time when processing information. Atthis stage in the program, however, theDELTIC had to be used in conjunction withsome other analyzer. Most of the other tech-niques are very slow and the outputs aresometimes difficult to interpret. In all ofthese techniques the problem of wow andflutter of the recorder was the limiting fea-ture.

10

The DELTIC Spectrum Analyzer

DEL TIC is a new data-processing technique

for the analysis of CTFM sonar returns. TheDELTIC equipment performs frequency analy-

sis by a sampling procedure and presents anoutput which is equivalent to the 100 -channelfilter bank analyzer typically used in CTFMsonar equipment. The DELTIC technique is

particularly advantageous for low-frequency,high-resolution analysis in real time. Theanalyzer has an interesting and useful feature

of storage and repeated analysis of data, either

automatically or by operator selection.

The conventional 100 -channel, multifilter

analyzer is generally designed with detectorsand envelope filters following each of the 100

filters. The purpose of the envelope filters is

to improve the signal-to-reverberation ratio

normally encountered in sonar equipment. Thetinne constants involved in the envelope net-

work depend on the scan rate of the sonarhydrophone as it passes a point reflector.

Envelope filtering should not be used in

searching for potential fish target signaturesbecause no restriction should be placed uponthe naturalness of the processed target signals

and the fish motion with respect to the scanningrate of the sonar set. To eliminate such re-

strictions a multichannel analyzer without

envelope filters would be required, and, there-fore, no provision could be made for minimizingthe unwanted reverberation. A multichannelanalyzer without envelope filtering would passa considerable degree of background noise orreverberation that would be presented on the

visual display and degrade the presentation.The DELTIC spectrum analyzer operates on

a totally different principle and can cope withthe variations in time responses expected for

the fish-type targets. The mechanism of the

reduction of the reverberation in such ananalyzer is considerably different from that

used in the conventional multichannel filter

analyzer and the equipment therefore can beused to examine fish-type spectra signatureswithout compromising perfornnance. TheDELTIC equipment operates with a wider fre-quency or range gate than that used for ultimateanalysis to sense the level of backgroundreverberation and apply appropriate compen-sation to the analysis capability of the DELTIC.The DELTIC also provides a "holding" featureby which interesting data can be retained, asdesired, for repeated display. In addition, theDELTIC analyzer has been designed to bedirectly interchangeable with the conventional100 -channel, multifilter analyzer in Strazasonar systems. This arrangement pernnits asystem to be used interchangeably as a sonaror as a Doppler analyzer.

Kay Vibralyzer and DELTICThe Kay vibralyzer was used for the analysis

of selected portions of the CTFM data wherepermanent records were needed. It was the

permanent output of the Kay vibralyzer that

first provided measurable proof of the surfaceand bottom reflections which were indicated

in the DELTIC display.About 5 minutes are required for the analysis

of 20 seconds of taped information. Thisrelation indicates the tremendous amount of

time required for analyzing test recordings;therefore, a technique had to be found whichcould be used for searching rapidly to isolate

portions of the tape for later analysis with the

Kay vibralyzer. The DELTIC was used for the

search, and excerpt tapes were then producedfor fine analysis and permanent record with

the vibralyzer. Figure 7 indicates a typical

^i! r » i '-

•Viw —

i<0o'r-

> < i

'i- ; .

-14L_

.— sKtuiiui = J i « » 5 i J

. !T' •, t I i i

•

y f

c^^f i((L-

Ut)

D— d^oO '>(.&]&

Figure 7.—Sample output from Kay vibralyzer showing double return from SCUBA diver at Mission Bay

Yacht Club site.

11

spectrum output from the Kay vibralyzer. Thehigh sweep rate of the CTFM sonar operatingat short range indicates clearly the rate of

bursts of data. It is possible that A j andAf could produce double signals suchas thoseindicated, but for the sweep rate of 34 ms.and the range of the diver, this double signal

is impossible. The first return at 1,800 Hzoccurs for the diver at a range of 14 feet

(4.3 m.) which agrees with his observed po-sition. The second signal can only be accountedfor by reverberation between surface andbottom.

U.S. Naval Ordnance Test Station

A trip to USNOTS, Pasadena, at the invitation

of J. Vetter and R. Davis, resulted in ananalysis of a portion of one of the Dopplertape recordings made at the Mission BayYacht Club pier. A lengthy discussion withthe personnel at USNOTS indicated that

extractions of signals from the returns wouldnot be easy and would require considerableprocess analysis. A day was spent at USNOTSusing their playback equipment and recordingsof fish returns from Mission Bay. The timerequired to analyze these portions of the tapebecame uneconomical.

Slow Scan Analysis Technique

A rapid means of analysis therefore wasattempted with a narrow band Hewlett-Packardlow-frequency wave analyzer and the testconfiguration indicated in figure 8. This tech-nique made use of intensity nnodulation of a

CRT (Cathode Ray Tube) and produced animage suitable for photographic integration.For the first time intensity-modulated signalswere integrated to improve the signal-to-noiseratio of the recorded data. The dynamic rangeof this technique was limited by the CRT,and therefore the output was confused by largenoise impulses. A sample of a typical photo-graphic output is included in figure 9 to indicate

AMPEX

RECORDER

351

the output obtainedby this particular technique.This technique requires 30 minutes to generatea completed presentation. Although the tinne

seems long for analyzing only 5 seconds of a

tape recording, it was short compared to thetime necessary for hand reducing the tapeinformation obtained from USNOTS. Theamount of tape that could be analyzed in thisfashion was certainly limited, and a quicksearch method was still needed for locatingsections of tape of interest. In the photographsobtained with this technique, certain indicatedanomalies were later shown to be produced bythe wow and flutter of the tape recorder in

both the recording and playback mode. Stepswere taken to remove the playback wow andflutter and to. reduce the wow and flutter pro-duced in recording. Careful observation of thefigure shows faint sinusoidal curves passingthrough the noise. These signals appear to

correlate with the motion of the yellowtailspecimen used in the Mission Bay Yacht Clubtests.

3-D Technique

The final technique that was used for theproduction of individual photographs necessaryfor the construction of a three-dimensionalmodel is shown in block diagram (fig. 10) andindicates the simplicity of the test setup.

The technique is reasonably good, but, asbefore, the time required is 15 seconds perpicture. Figure 11 shows a sample of one ofthe photographs produced from the CRT usedin this approach.

The model, composed of 25 individual photo-graphs such as that in figure 11, is shown infigure 12. This model of one film "take" atCatalina contains information on the activityof the fish in the area of the chumming plat-form. The correlation between the activity offish in the test photographs and the model is

reasonable. The density of activity was toogreat to isolate individual specimens as in theMission Bay tests, but some individual corre-lations could be seen in the model constructed.

DETECTOR

Figure 12.— Photograph of 3-dlmenslonal model used to establish correlation between motion pictures andrecorded data.

Three models were constructed to demonstratethe Doppler correlation characteristics be-tween data and the motion pictures.

CONCLUSIONS FROM FEASIBILITY STUDY

The results of the work accomplished clearlyindicate the suitability of CTFM sonar for

ranging on fish specimens. This fact wasdetermined early in the program.

The remainder of the effort and time wasgiven to generation of techniques for presen-tation of the weak Doppler return from indi-

vidual fish targets. The results obtained werenot decisive but indicated that with a greaternumber of specimens the signal-to-noise ratio

improved and the data extraction techniquesrelaxed. The application of a low-frequencyDELTIC to the analysis of the Doppler returnsproduced a simple presentation. The techniquesand circuitry that were developed overcamemany difficult problems.

In view of the higher target strengths of

schools of fish, it is clear that reasonableranges of detection can be achieved with this

type of sonar (e.g., 1,000 yards--914,4 m.).The coherence of the dynamics of the schooland increased signal strength provide addi-tional significant analysis information.

SHIPBOARD TARGET CLASSIFICATIONEXPERIMENTS

The initial attempts to detect Doppler effects

caused by fish showed that the technicalproblems could be solved. How well theorycould be reduced to practice was to be testedin the next, the at-sea, phase. The equipmentused in Phase II was modified and packagedfor shipboard use. We decided that "realtime"signal analysis was necessary and incorporated

a DELTIC analyzer in the system; however,we also provided an instrumentation recordercapable of storing target echo signals for

laboratory analysis should the need arise.

Equipment

Our target-classification sonar system wasa modification of a Straza Electronics 500series CTFM sonar (table 3). CTFM sonarhas several desirable features for use onpelagic fishes. A wide-beam projector con-tinuously fills a large volume of water withsound, the frequency of which is repeatedlyvaried from high to low in a saw-tooth manner.A narrow-beam hydrophone rapidly scans the

insonified volume (25°/second in our sonar).All targets in the projector beam returnechoes at a frequency that differs with time(or range) from the projected frequency.Mixing of the transmitted and received fre-quencies produces a difference frequency cor-responding to a range. This continuous proc-ess gives a much faster information rate

than is possible with a conventional pulsedsonar systenn. This high information rate

makes it possible to maintain contact withfast-moving schools. Each target is a sourceof continuously reflected sound. The hydro-phone scans each target for a time equivalentto several pulse lengths of a pulsed sonar; as

a result the target returns are averaged. Suchsignal-averaging of short-term fluctuations in

echo strength increases the likelihood of de-tection of any particular target.

Because range information was obtainedfrom a difference frequency, CTFM sonaralready has some of the circuit design that

was required for resolution of the Dopplereffect. An addition of a second mode of opera-tion CTD (Continuous Transmission Doppler),that used most of the CTFM circuits plus a

14

Table 3.—Specifications of SM506 sonar system

ProjectorSource frequencySource levelHorizontal beam widthVertical beam width

Receiving hydrophoneHorizontal beajn widthVertical beam widthOpen circuit sensitivy

CTPM ranges

Horizontal scan rate

Presentation CTM7-inch plan position indicator scope5 -inch "A" type display on CRT

Audio

Presentation CTD7-inch relative bearing display5-inch frequency spectrum displayAudio

60-75 kHz, 70 kHz in CTD mode+95 db re it bar/volt 1 yd,20°20°

4°

10°, 20°- 75 db re 1 volt/// bar

100, 200, 4-00, and 800 yds.

25°/second on 100, 200, and 400 yds.10°/second on 800 yds.

Data storageVideo, audio, range, servo, and mode information to multiplexer and Ampex

Model SP-300 tape recorder. Playback through system.

frequency analyzer of higher resolution, per-mitted our sonar to be used to locate and also

classify a target. The system componentswere packaged to mount on vessels of con-venience; the underwater unit was mountedon an over-the-side column.

Methods

When a fish school was located, the CTFMmode was used to position the boat 100 to 200yards (91.4 to 182.9 m.) from the school. If

the school was moving we tried to match its

velocity and direction of movement with the

vessel. When the school was positioned, the

sonar scan was locked and the system placedin the CTD mode. In this mode a continuous70 kHz tone was transmitted fronn the sonarprojector. The echo of this tone plus Dopplerinfornnation was received by the hydrophone,mixed and filtered to remove the portioncorresponding to 70 kHz, and fed to the

spectrum analyzer. The spectrum analyzer hada DELTIC to raise the Doppler-frequencyinformation by factors of 2 or 4 by lO"' andallow frequency resolutions of 3.33 Hz or 1.67

Hz over a band of ± 600 Hz or + 300 Hz. TheDoppler spectrum could be displayed on a

CRT and photographed or recorded on magnetic

tape, or the same information could be storedand displayed repeatedly on the CRT by locking

the DELTIC.The incoming signal was sampled by logic

circuitry to convert Doppler-frequency infor-

mation into digital form. This digital

information was loaded into the delay line at

the rate of 100 ^ sec./bit when the + 600 HzDoppler band was used and at 200 // sec./bit

when the t 300 Hz band* was used. Since the

DELTIC could store 3,000 bits of information,the initial loading time of the analyzer was 0.3

sec. or 0.6 sec., depending on the resolution

selected. After the line was loaded, newDoppler information could be added and old

information dropped at the rate of about onebit/150 n sec. The output of the analyzer thenwas analogous to a photographic time exposureof the Doppler data received in a 0.3- or0.6-sec. interval. The analyzed Doppler infor-

mation was presented as video on the CRTat 5 sweeps per second. Frequency wasdisplayed along the horizontal tinne line withcenter frequency (no Doppler) at the centerof the scope, up -Doppler to the left, anddown-Doppler to the right (fig. 13). Vertical deflection

depends on the number of times a particularfrequency was present during the load time of

the delay line. A single frequency was shownas a single high spike, whereas multiple

15

®-««vj 9.-9

Figure 13(a).--Typlcal narrow-band Doppler from a small metal triplane at 10 yards (9.1 m.): no motion, large

spike at Doppler, horizontal scale 60 Hz/div, vertical scale relative; (b) motion of triplane toward hydrophone,

spike displaced to left; (c) motion away from hydrophone, spike displaced to right.

16

frequencies were displayed as many spikes of

lower amplitude.In practice we felt that the sonar

information--f i s h swimming speed frommatching the ship speed to the speed of theschool and Doppler shifts caused by tail beatsand body motion- -would result in some approx-imation of fish size and type. (For moredetailed analysis of these relations and somemore examples of results see Hester, 1967.)

Results of Shipboard Tests

Over a 2-year period we tested our equip-ment under a variety of conditions on severalspecies of fish. We encountered two majordifficulties to the obtaining of Doppler recordsfrom fish schools.

The first difficulty was sea-noise Dopplereffect from surface reflections. Theover-the-side transducer array was 10 feet (3.0 m.)

below the surface. This shallow depth combinedwith the 10° vertical beam width of the hydro-phone allows surface reflections to be detectedreadily. To receive a Doppler recordfrom fish,sea conditions must be extrennely favorable--no wind ripples or short-period swell. Suchconditions do occur, particularly in the earlymorning, off the Pacific coast, but the oc-currence of fish schools and "workable" con-ditions together are rare. It was soon apparentthat commercial applications of Doppler equip-ment would require a more complex signal-processing procedure than was originally ex-pected.

The second major difficulty arose fronn theirregular behavior of the large schoolingfishes such as tunas. These fishes, becauseof their streannlined shape, are detectablewith our sonar only when viewed from theside (fig. 14), When schooling, scombroidsusually travel at a good rate (3 to 5 knots)

Figure 14. --Target echo video from a fresh 492-mm. black skipjack ( Euthynnus Uneatus ) suspended 5 m. from the

sonar dome. The sonar gain was set to saturate the receiver for the target's side aspect (a). The echo video is

markedly reduced when the target is rotated head (b) or tail (c) to the beam.

17

and make capricious course changes. Wefound it exceedingly difficult to nnaintain con-tact with these schools. Even if contact wasnot broken the schools had to be kept within100 yards (91.4 m.) of the vessel for Dopplerdetection, and at these close ranges theyusually outmaneuvered us.

Because of these difficulties we decided to

see if the complete target signal, noise Dopplereffect and target Doppler effect could beclassified by computer. A contract was madewith Scope Inc., Falls Church, Va., to exploreacoustic methods of classifying characteristicsof schooling fish. Specifically, the work wasconcerned with establishing the feasibility of

using adaptive pattern-recognition processesin the automatic classification of Straza CTFM/CTD sonar returns.

Data were collected aboard the Bureau of

Commercial Fisheries vessel. Miss Behavior ,

off the coast of southern California, The audiooutputs of the CTFM and CTD receivers, asselected in the DELTIC analyzer, were re-corded on one channel of the associated AmpexSP300 magnetic tape recorder; a voice annota-tion was recorded on another channel. When atarget searched for in the 200-yard (182.9 m.)and 400-yard (365.8 m.) CTFM mode waslocated, it was recorded with voice annotationof estimated target size, species, approximategeographical location of the vessel, range-to-target, and sonar mode of operation. Upontarget detection, the "Scan Rev" control wasmanually activated to reduce the scan from± 45° or 22.5° to about t one-half of the anglesubtended by the target. This procedure allowedmore nearly constant contact with the targetwhile permitting it to be tracked in azimuth.

The recordings thus made were subsequentlydubbed twice to make them compatible withscope's magnetic-tape reproduction facili-

ties. The resulting tape contained six targetsand five ambient recordings, totaling about 18minutes of usable data. Table 4 shows a de-tailed breakdown of the data.

It became apparent after the initialprocess-ing of the magnetic tapes that the limited database obtained was insufficient for the analysisrequired to determine the feasibility of usingpattern-recognition processes in the automaticclassification of sonar returns. We did under-take analysis, however, with the data availableto ascertain whether the sonar echo receivedwhen a target is present contains anomaliesthat are not in the echoes received during the

recording of ambient conditions. Further, it

was intended that, if such anonrialies werefound, the analysis would reveal whether theyhave characteristics that might be used in

classifying the target as to size and species.The decision to analyze for spectral content

the data available in the initial study wasdictated by two factors. The first considerationwas that the difference frequency in the CTFMmode is the only expression of the information

available in that mode--target range and rela-tive target motion. The latter produces aDoppler shift of the FM signal, resulting in a

modulation of the difference frequency. Thesecond consideration was that (as indicated byan exannination of the physics of the situation)

the information present should be in the formof spectral/time distributions. Furthermore,this analysis could fortunately be produced byinstrumentation readily available,Sonagrams were made of the data on a

Sonagraph 606 1A manufactured by the KayElectric Company. This spectrunn analyzerproduces a frequency versus time versusamplitude (intensity) plot in the 85 Hz to 8

kHz range. When a target was known to bepresent, a signal occasionally appeared at a

frequency corresponding to a range approxi-nnating that of the target--when the targetrange was known. Targets never appeared in

more than one -third of the sweeps becauseof the sector -scanning technique of the sonar.

In both the CTFM and CTD modes and at

all ranges of the CTFM mode a signal appearscontinuously in the 5 to 5.5 kHz range. Attennptsto correlate the signal with a known target ora natural obstruction such as the ocean floorproved fruitless.

Further analysis was made on a GeneralRadio Wave Analyzer Model 1900A and re-corded on a General Radio Graphic LevelRecorder Model 1521B. The frequency-versus

-

amplitude plot produced disclosed a different

frequency signal with a value approximatingthat of the range of the target. Once again,the unidentified signal appeared, this time at

5.3 kHz.Signals were detected with both the sona-

graph and wave analyzer at frequencies ap-proximating those of target range. The signalsalso appeared at the times when a target wasbelieved to be present, but the detected signaland the target could not be correlated otherthan in range and time. Classification of thetarget by species was impossible, and it doesnot appear practical to look for more informa-tive features until we obtain a larger andmore comprehensive data base.

RECOMMENDATIONS

To date the sonar experiments have giveninteresting but inconclusive results. The editorbelieves that the prospects for commercial-fishing applications of Doppler target classi-fication are at this time financially impractical.The usefulness of the Doppler work in researchis less clear. The problem of identifying

subsurface sonar contacts for population andbehavior studies is yet unsolved. It is too

early to say whether acoustic methods --high-resolution, resonant studies or Dopplereffects --alone or in combination will provemore satisfactory than other methods such as

18

Table <+.--Log of targets observed

Target no. andlength of

observation

trawling or optics. Work is progressing in

all of these fields but this Straza unit is

apparently the only sea-going Doppler unit

under investigation. Soundness of its con-tinued use on surface targets is questionable.Experience suggests that the direction of the

research be changed. The difficulties causedby sea return suggest that the sonar shouldconcentrate on deeper targets. Since body-motion Doppler effect is aspect-dependent,that is, it is strong only on the horizontalplane, the transducer array must be mountedon a subsurface vehicle so that it can operateat the same depth as the target. The problemsarising from trying to maintain contact withrapidly moving surface schools may not be as

great with deep targets since some studies

show that these targets generally are nearlystationary. It is most likely, however, that

these deep targets are snnall organisms--which means that work on the big scombroids,the original goal of these experiments, will

have to be deemphasized. In addition we shouldhave a combination of laboratory studies of

fish locomotion along with the future Dopplerwork.

3. The at-sea experiments showed that

under certain conditions, Doppler informationcould be obtained fronn near-surface fish

schools. This information was not successfullyused in target classification, however. Themajor difficulties with the at-sea phase werecaused by sea return and the erratic behaviorof near-surface schools. Recommendations for

further studies are aimed at overcoming thesedifficulties. The recommendations includemounting the sonar transducers on an under-water vehicle and working with deeper, slow-moving targets.

ACKNOWLEDGMENTS

The editor acknowledges the cooperation of

the U.S. Navy for the loan and subsequenttransfer of our sonar boat, the engineeringstaffs of the contractors for their outstandingperformance and reports, and the DireccionGeneral de Pesca e Industrias Conexas, De-partmento de Estudios Biologicos Pesqueros,for permission to enter and work in the

territorial waters of Mexico,

SUMMARY

Three phases in the development of a Dopplertarget classification sonar are reported:

1. Target strengths of subject species weredetermined to permit the estimation of target-classification sonar equipment: frequency,source level, hydrophone and receiver sensi-tivity, and scan rate,

2. Feasibility studies were made to deter-mine whether or not Doppler effects from fish

body-motion could be detected. These studies

were sufficiently promising that a shipboardCTFM/Doppler sonar was constructed.

LITERATURE CITED

HESTER, FRANK J,

1967, Identification of biological sonar tar-

gets from body-motion Doppler shifts.

In Wm. N. Tavolga (editor). Marinebio-acoustics, pp. 59-74. PergannonPress, Oxford and N.Y.

INTER-AMERICAN TROPICAL TUNA COM-MISSION.1960. Annual reportfor the year 1960. Inter-

Amer. Trop. Tuna Comm. 183 pp.1961. Annual report for the year 1961. Inter-

Amer. Trop. Tuna Comm. 171 pp.

20

Acoustic Target Strength of Several Species of Fish

By

H. W. VOLBERG^Vice President-Manager

Straza ElectronicsDivision of Straza Industries

El Cajon, Calif. 92021

ABSTRACT

To design fish-finding sonar equipment it is necessary to have information about

target strengths of fish. This study was made principally to determine the target

strength of tunas at several acoustic frequencies. In addition, measurements weremade on other living, dead, fresh, and frozen fresh-water and salt-water fishes,

Sonne without swim bladders.

INTRODUCTION

The use of sound waves to locate objects in

the sea has been refined over the last fewdecades. Most of the impetus for the develop-ment of equipment and techniques came fromthe military. Peaceful application, especially of

scanning sonar equipment, often lagged far be-hind the potential of the military capabilities.

Many of the features necessary for sonar sys-tem design are available from the military;

however, the most important one for fisherywork, the target strength of a fish, was not

available from the military.The problems involved in fishery work are

complicated by such variables as the presenceor absence of a swim bladder in the fish, the

behavior of the species (schooling or solitary),

the size of the individuals, the expected orien-tation of the fish to the sound beam (vertical orhorizontal echo ranging), and the relation of

fish target strength to the sound frequencyused. Recently, European experiments haveyielded data on several species. These in-

vestigations (Harden-Jones and Pearce, 1958;

Midttun and Hoff, 1962) were chiefly concernedwith determining size/target strength relation

and the contribution of the swim bladder to

target strength. Gushing (1964) combined the

other European data with his to illustrate a

fish length-target strength relation applicablefor middle frequencies (near 30 kHz), The

species reported have swim bladders; some of

Cushing's fish had artificial swim bladdersplaced inside their bodies.

The primary demand in American fisherysonar work comes from the groups working onthe fast-swimming, often pelagic species suchas scombroids and salmonoids. Differences in

target strength between side and dorsal aspectire of interest as is, for example, the differ-

ence in target strength between a yellowfin

tuna ( Thunnus albacares ), a species with a

swin bladder, and a skipjack tuna (Katsuwonus

pelamis ), a species without one.

In 1963, Straza Electronics was invited to

submit a proposal for a high- resolution sonarsystenn for fishery research. To calculate

energy requirements and receiver sensitivity

values for the proposed system it was neces-sary to know something about the target char-acteristics of the species to be studied (chiefly

tunas). Also, the system envisioned would bemultifrequency, using frequency modulatedbands, one of several tens of kilo Hertz andone of several hundreds of kilo Hertz. Thisreport summarizes the tests nnade to deter-mine target strengths for several species of

fish at various frequencies and for variouspositions of orientation. Differences of

target strength between living, fresh, andfrozen specimens and the contribution of

the swim bladder to the sound reflectiveproperties of some fishes also were investi-gated.

^This work was sponsored and funded in 1963 by Straza

Electronics, with the cooperation of the U.S. Navy Elec-

tronics Laboratory and the Bureau of Commercial Fish-

eries.

2 Present address: Consultant In Acoustics, 5729 Gen-

natte Avenue, La Mesa, Calif. 92041.

21

FACILITIES, EQUIPMENT, AND METHODS

The equipment and facilities of the U.S. NavyElectronics Laboratory, Sweetwater Dam fa-

cility, San Diego, Calif., were used to deter-mine the target strength of the various fishes.

Target strength is defined as 10 log the ratio

of the acoustic reflectance of the specimen re-

ferred to a perfectly reflecting sphere of 4 m.diameter at a range of 1 m. It is expressed in

db (decibels). Figure 1 shows a block diagramof the test setup used to nnake our measure-ments of target strength. The fish target wasplaced appropriately at either 6 or 12 m., de-pending on the frequency, from the trasmit-receive transducer so that near-field effects

were not evident. A 1 -millisecond pulse wastransmitted at the carrier frequencies of 20,

40, 50 or 280 kHz. At an appropriate time, the

receiver gate was opened, to permit the echopulse to be recorded in amplitude on the polarplot equipment. The pulse repetition rate wasadjusted to minimize the interference in the

water. The fish target was mounted on mono-filament rigging which was rotated by meansof a servo system at about l°/second. Therecorder response was capable of 50 db/second,20 db/half second, and 10 db/third of a second.

To support the fish specimen properly, the

rigging took the shape of an inverted "A" with

a weight at the bottom (fig. 2). Wire gave

r "

acoustic reflections which had too great am-plitude, especially at 280 kHz, for the typesof tests anticipated. Monofilament nylon af-forded an improvement of some 15 db overthe wire rigging at 280 kHz and was usedthroughout the tests.

TESTS

Previous studies (Harden-Jones and Pearce,1958; Midttun and Hoff, 1962) show that theswim bladder in fishes is an important con-tributor to their acoustic cross-section. Sinceskipjack tuna (Katsuwonus pelamis ) lack aswim bladder yet are of considerable com-mercial importance, they were tested for com-parison with fishes such as the yellowtail

( Seriola dorsalis ) that have swim bladders.

Fish Without Swim Bladders

To illustrate the nature of the directivitypatterns obtained, two figures are shown for askipjack tuna. This tuna was 60 cm. long,weighed 5.5 kg., and initially had been chilledbut not frozen for 48 hours. Measurementswere made at two frequencies --40 kHz and280 kHz. Figures 3 and 4 show patterns ob-tained when the dorsal fin was in the verticalplane and the body rotated in the horizontal

SKIPJACK ICh.lladI

12 poundi 60 cm long

DItlCTIVITT MTTIRN

DIRiCTIVITY PATTERN

= 270'

SKIPJACK iCh.lledl

12 poundv 60 cm

F.eqcency 40 liMc =270" Bolole 6 Depth 3 90 m.leti

Tetr D.ilonce 6 m.rc. Tempefoiu-e 2rC Scoie lOdboftH

Figure 3.—Directivity pattern for a 60 cm. skipjack tuna

rotated in the horizontal plane. The test frequency was40 kHz. For this and following figures the target range

was 6 m., the temperature 21° C, and the test depth

3.9 m. Each ring represents a 10 db decrement from the

outer -20 db ring.

66 em FROZEN

Figure 5.— Rectilinear plotof directivity patterns of 40 kHz(left side only) of four skipjack tuna. See text for ex-

planation.

270 290 310 330 350 10

BACK 0°

d - 90° ROTATE

70 cm LONG - 40 kHi.

60 cm LONG - 40 kHs.

60 cm LONG - 280 kHz.

70 cm LONG -280 kHi.

anatomical parts of a skipjack tuna were esti-mated by the progressive dissection of a 7 2 -cm.specimen. First a reference pattern was run;then patterns with the opercle and preopercleremoved from one side--from both sides--thecorselet removed- -pectoral and pelvic gridlesremoved--head, tail, and viscera removed--flesh from one side removed- -and flesh fromboth sides removed. These tests indicated thatany large cross-sectional area of any part ofthe fish showed a substantial echo with respectto the target strength of the intact fish. At280 kHz no significant difference existed in thebeam aspect target strength when only the skel-eton of the skipjack tuna was used minus headand tail, compared to an intact fish.

Fish With Swim Bladders

A frozen yellowtail, 80 cm. long, was used ina sTies of tests on fish with swim bladders,f igure 7 shows a series of directivitypatternsfrom the 80 -cm. -long yellowtail suspended noseup, tail down, so that the back and sides of thefish could be examined. The top two directivitypatterns are for 50 kHz--first with the fishintact and second with the swim bladder punc-tured and flooded. The bottom two patterns arefor 280 kHz under like conditions. The 50 kHzfrequency was selected because the effect ofthe swim bladder for this specimen appearedmore pronounced there than at any other fre-quency.

Figure 6.— Directivity patterns for two skipjack tuna

suspended nose up, tail down and rotated about the long

axis.

The irregular nature of the directivity pat-terns showed the fish to be a highly complexacoustic target. The surface of a fish andvarious internal structures both are believedto contribute to the reflectivity of the wholeanimal. The relative contribution of various

A 50 kH>,

B 50 kHz - BlADDEfi PUNCTURED & FLOODED

C 280 kH<

D 280 kHz - BLADDER PUNCTURED & FLOODED

Figure 7.— Directivity pattern for a yellowtail suspendednose up, tail down and rotated about the long axis. Tests

were made with and without the swim bladder intact.

24

The plot in part (A) of figure 7 shows a -27db target strength near the back portion of thefish. This area is located along the abscissaof the graph at about 0°. In the same area, thereturn in part (B) was markedly reduced whenthe bladder was punctured and flooded. Bywayof comparison, the higher frequency patternsin parts (C) and (D) show no appreciable dif-

ferences. The influence of the swim bladderat the lower frequencies might be due to

lessened attenuation by the flesh at thosefrequencies and a consequent increase in theresponse from the swim bladder. For fish of

the size of the yellowtail tested, and assumingtypical flesh attenuation to be in the order of

0.6 db/cm. at 280 kHz, a difference of some7 db can be attributed to flesh attenuation(Goldman and Heuter, 1957). This increasedattenuation at the higher frequencies couldmask the influence of the swim bladder. It is

also conceivable that some misalignment oc-curred during the repeated aspect angles at

this high frequency.A similar experiment was run with yellowfin

tuna ( Thunnus albacares ) at a frequency of

40 kHz. In this test, the fish was suspendedboth in normal swimnning position and nose up,tail down. A typical pattern for the yellowfintuna is given in figure 8. Figure 9 shows aseries of directivity patterns for the 73-cm,-long yellowfin tuna for 20 kHz and 40 kHz. Acomparison is made between an intact fish andthe same fish with its swim bladder puncturedand flooded. For this observation (head up,tail down) the back portion of the fish was

DIHECTIVITr PATTIBN

= 270YELIOWFIN TUNAIFfoirnJ 73 em long

* = 270 Rol

Figure 8.—Typical 40 kHz directivity pattern for a yellow-

fin tuna suspended In the normal swimming position.

9 • 90' ROTATE * 73 cm LONG

A 20 kHi

B: 20 kHz - BLADDER PUNCTURED & FLOODEDC: 40 kH!

D 40 kHi - BLADDER PUNCTURED & FLOODED

Figure 9.— Directivity pattern for a yellowfin tuna sus-

pended nose up, tall down and rotated about the long

axis. Tests were made with and without the swim bladder

Intact.

viewed. At 20 kHz, difference in response be-tween the intact fish and the fish after theswim bladder had been flooded was not appre-ciable. At 40 kHz, however, the return de-creased significantly when the swim bladderwas flooded. With the above exceptions, all

the tests showed that in general the targetstrength of yellowfin tuna was comparableto that of skipjack tuna and yellowtail of ap-proximately the same size.

Additional test were made with largemouthbass (Micropterus salmoides ) and white crappie

(Pomoxis annularis ), both alive and dead. Whenthe bass and crappie were alive, the gill

movement was readily observable on the plotsof the directivity pattern. Difference betweentarget strength in live and dead fish was scantor nil. The target strengths were 15 db lowerthan those of the larger tunas due to the smallersize of the fish (largemouth bass 40 cm,, whitecrappie 32 cm,). Deflating the bladders de-creased the target strength 3 to 7 db,

RELATION OF PRESENT TO EARLIERFINDINGS

It is evident that the target strength of a

fish depends largely on its size. Only in someaspects, especially when the fish is viewedfrom above, does the swim bladder contributesignificantly to the strength of the echo. Thecontribution of the swim bladder rapidly de-creases with increased frequency, apparentlybecause of an increase in attenuation by flesh

25

with rise in frequency. The observations ofHarden-Jones and Pearce (1958) and Midttunand Hoff (1962) showed an increase in targetstrength and directionally with fish size up to

about 70 cm. At this point the thickness of thetissue between the swim bladder and the sur-face of the back of the fish is sufficiently greatto begin to nnask the contribution of the bladder.The earlier results are combined with some of

ours in figure 10. The abscissa is in units of

20 log fish length to put the contribution of sizeinto logarithmic units. Target strengths ofboth side and back aspect are given for ourlarger fishes. In open water these species willusually be detected at some slant range andtheir target strength will lie somewhere be-tween the values given for side and top. Thedecreases in target strength due to floodingof the swim bladder are indicated by thevertical arrows.

LENGTH dF FISH (cm.)

40 50 60 70 80 90 100

32 31 36 38

20 LOG LENGTH

SKIPJACK Tu^A, SIDE ASPECT

rELLOWFlN TUNA, SIDE ASPECT

YELLOWTAIL. BACK ASPECT

YELLOWFIN TUNA, BACK ASPECT

CRAPPIE,

BASS.

SKIPJACK TU

YELLOWTAIL.

YELLOWTAIL. (C

YELLOWFIN TU

CRAPPIE,

BASS,

lA, BACK ASPECT

SIDE ASPECT

Ibockl, BLADDER FLOODED

:h|, BLADDER FLOODEC

1 perch] JONES a PEARCE 1959

2 COO, BACK ASPECT3 COALFISH, BACK ASPECT

MIDTTUN a HOFF

Figure 10.—Comparison of target strength measurements

made in this experiment with those obtained by other in-

vestigators. All measurements were made between 20

and 40 kHz.

SUMMARY

1. The acoustic target strength for skipjackand yellowfin tunas, yellowtail, white crappie,and largemouth bass were measured for dif-

ferent orientations of the fish at frequenciesfrom 20 kHz to 280 kHz.

2. The differences in target strength amongliving, dead, chilled, and frozen fish werenegligible,

3. The contribution of the swim bladder to

target strength decreased as size of the fishand test frequency increased.

4. With the exception of the contribution of

the swim bladder, target strength was inde-pendent of frequency.

ACKNOWLEDGMENTS

J. Roshon and A. Huntly of the U.S. NavyElectronics Laboratory, San Diego, providedtechnical assistance. Personnel of the Bureauof Commercial Fisheries Biological Labora-tory, Honolulu, and the Bureau of CommercialFisheries Fishery-Oceanography Center, LaJoUa, assisted by supplying test specimens.

LITERATURE CITED

CUSHING, D, H.1964. Cons. Perma. Int. Explor. Mer, Con-

tributions to Symposium 1963 on theMeasurement of Abundance of FishStocks. J. A. Gulland[ed.], 34: 190-195.

GOLDMAN, D, E., and T. F. HEUTER,1957, Errata: Tabular data of velocity and

absorption of high frequency sound in

mammalian tissues. J. Acoust. Soc. of

Amer. 29: 655,HARDEN-JONES, F. R,, and G. PEARCE.

1958. Acoustic reflection experiments withperch ( Perca fluviatilis Linn.) to de-termine the proportion of the echo re-turned by the swim bladder. J. Exp.Biol., 35: 437-450,

MIDTTUN, L., and I. HOFF.1962, Measurements of the reflection of

sound by fish. Rep, Norw, Fish, Mar,Invest. 13(3): 18 pp.

MS. #1997

26

GPO 888.770

MBL WHOI Llbfarv - Serials

5 WHSE 01794

As the Nation's principal conservation agency, the Depart-

ment of the Interior has basic responsibilities for water,

fish, wildlife, mineral, land, park, and recreational re-

sources. Indian and Territorial affairs are other major

concerns of America's "Department of Natural Resources."

The Department works to assure the wisest choice in

managing all our resources so each will make its full

contribution to a better United States — now and in the future.

UNITED STATESDEPARTMENT OF THE INTERIOR

U.S. FISH AND WILDLIFE SERVICEBUREAU OF COMMERCIAL FISHERIES

WASHINGTON, D.C. 20240POSTAGE AND FEES PAID

U.S. DEPARTMENT OF THE INTERIOROFFICIAL BUSINESS

Return this sheet to above address, if you do

NOT wish to receive this material| 1

, orif

change of address is needed| |

(indicate

change including 3P Code).