deteção de gelidium sonar

7/26/2019 Deteo de Gelidium Sonar

1/12

Bol. Inst. Esp. Oceanogr. 15 (1-4). 1999: 325-336 325

Acoustic methods for Gelidiumseaweed detectionA. C. Molero and R. Carb

Instituto de Acstica, CSIC. Serrano, 144. 28006 Madrid, Spain

Received September 1997. Accepted April 1998.

ABSTRACT

The reflection and absorption coefficients of a thick layer of Gelidium sesquipedale(Clem.Born. et Thur.) seaweed covering a sandy bottom were determined in a laboratory tank. As a re-sult of the reduced vertical cross-section and the high water content of this seaweed, its acousti-cal impedance is very similar to water impedance, and the target strength of each individual sea-

weed frond is very weak. The usual frequency range for detection of marine life, 100-500 kHz,has been used. The average value of the bottom scattering strength was found to be between26 dB and 34 dB in the frequency band used.

Key words: Underwater acoustics, bioacoustics, back-scattering.

RESUMEN

Mtodos acsticos para la deteccin de Gelidium

Se han realizado mediciones del coeficiente de reflexin y absorcin de una capa espesa de algas deGelidium sesquipedale (Clem. Born. et Thur.), sobre un fondo de arena, dispuesto en un tanque de expe-riencias hidroacsticas. Como resultado de la reducida seccin transversal acstica de cada brote del alga, yde su alto contenido de agua, su impedancia acstica tiene un valor muy prximo a la impedancia acsti-ca del agua marina y, en consecuencia, el nivel de blanco de cada brote es muy dbil. El rango de frecuen-cias utilizado parte de 100 kHz, alcanzando 500 kHz. El valor promedio del nivel de difusin de fondo dela capa de alga de Gelidium encontrado vara entre 26 dB y 34 dB en el rango de frecuencias utilizado.

Palabras clave:Acstica subacutica, bioacstica, retrodifusin.

INTRODUCTION

To improve the detection possibilities of Gelidiumsesquipedale (Clem. Born. et Thur.) seaweed in theseabed, its acoustical behaviour has been investigat-ed under laboratory conditions, to determine thebasic acoustic properties: reflectivity, absorption andacoustic impedance, in the 100-500 kHz frequencyrange.

G. sesquipedale is a Rodofita seaweed (figure 1)which belongs to the Rhodophyceae class, Gelidiales

order, and its size varies from 2 mm to 45 cm, with auniaxial structure. They are finally shaped intogrouped clusters; from a very active rhizoid, a suc-cession of new portions of thellus begins to grow.This rhizoid attaches to a calcareous substrate, witha predominantly organic origin, through tiny hap-teron discs (Juanes and Borja, 1991).

Because of its high content in agar-agar, there isuniversal interest in this seaweed. Agar-agar is a

very complex polysaccharide, highly valued as athickener in the food industry. It is also used in bac-

Bol. Inst. Esp. Oceanogr. 15 (1-4). 1999: 325-336 INSTITUTO ESPAOL DE OCEANOGRAFAISSN: 0074-0195

Ministerio de Agricultura, Pesca y Alimentacin, 1999

BOLETN


2/12

teriological synthesis, and this latter application isanother source of interest in agar-agar. Still, not allagar-producing seaweed are useful in bacteriologi-cal agar production, only five of them, includingG. sesquipedale, are able to produce agar of suffi-cient quality.

The increase of microbiological synthesis tech-niques has made agar a highly interesting product,due to the universal lack of high-quality agar-pro-ducing seaweed. Its value on the international mar-ket is growing daily as new applications and sub-products derived from agar appear continuously.

Agarose, which is a sulphate-free colloid extractedfrom agar, commands a very high internationalmarket price.

On the Atlantic Iberian continental margin,large Gelidiumseaweed communities can be found,

especially on the Cantabrian continental shelf, fromFuenterraba to Cabo de Peas. There are also oth-er major Gelidiumseaweed populations, located onthe Galician coast from Camarias to La Guardia,and on the Portuguese shelf, especially near Porto,Setbal and Algarve (lvarez et al., 1989; Juanes andBorja, 1991). Spain probably has the highest-densi-ty seaweed community in the world.

It is noteworthy that one of the natural forms inwhich Gelidium can be found after loosing from theseabed is known as arribazn. This is a seaweed mass

no longer attached to a calcareous substrate, sothat it can move anywhere with the currents, float-ing in the water column. If the wind blows towardsthe coast, people on the beaches can collect it eas-ily, but whenever the wind blows in the opposite di-rection, the arribazn is transported towards lowdynamic regions, remaining there until its decom-position (in about 10 days).

The application of new acoustic detection tech-niques (i.e. sonar) in the finding of Gelidium sea-

weed masses, would enable us to make use of theseas natural resources, avoiding the loss of tons ofGelidium seaweed, whenever conditions on the con-tinental margin are inadequate for beach collection.

MATERIALS AND METHODS

Underwater acoustic detection requires knowled-ge of seaweeds main acoustic properties, e.g. reflec-tion coefficient, acoustic impedance and absorption.

These parameters have been quantified largelythrough detailed laboratory studies. The transducerprojects a pulse in a directional beam. The trans-mitted pulse of sound propagates through the wateraway from the transducer. It may encounter varioustargets, e.g. seaweed and the sand bottom. Thesetargets reflect or scatter the pulse, and some energyreturns towards the hydrophone, being recordedand processed by an A/D converter (figure 2).

In the present experiment, Gelidium seaweed (ina layer 20 cm thick) was placed in a water tank(1.06 m 0.71 m 0.90 m), taking special care toobtain a homogeneous volume concentrationthroughout the entire layer. Under the seaweed lay-er, a sandy bottom simulated natural conditions,

with a thickness of 11 cm, grain size < 2 mm and av-erage density 2.65 g/cm3.

Relevant parameters temperature, salinity andseaweed concentration were controlled through-out the experiments. Water in the tank with marinesalt in suspension reassembled ocean water, reach-ing a salinity value of s = 35 , and density r == 1.024 g/cm3. Air-bubbles attached to the seaweed

were also carefully eliminated.Three piezoelectric echo-sounders were used to

cover different frequency ranges, directivity andduration of the transmitted pulse. Table I shows theessential features of the echo-sounders.

A. C. Molero and R. Carb Acoustic methods for Gelidium seaweed detection

Bol. Inst. Esp. Oceanogr. 15 (1-4). 1999: 325-336326

Figure 1. Gelidium sesquipedaleseaweed


3/12

The directivity pattern, emitted pulse and fre-quency response of each transducer is shown in ap-

pendix I.As a receiver sensor, a Brel & Kjaer hydrophonewas used, placed by the transducer, both working asa mono-static system (appendix I also includes hy-drophone performances).

The combined echo-sounder/hydrophone x-yaxis movement was controlled automatically, withthe z axis remaining at a depth of at 2 cm deep.Three parallel trials, at y = 30 cm, y = 35 cm and

y = 40 cm, with 30 acquisition points (every cm),made up a total recording network of 90 points.

RESULTS

Echoes produced by the sandy bottom without sea-weed (a), by the seaweed layer (b), and by the sandybottom with seaweed above (c) were recorded at thethree frequencies studied: 100, 200 and 500 kHz.

Figure 3 presents an example of situations band c. The figure presents the signal taken by thehydrophone with the three different echo-sounders transmitting at 100, 200 and 500 kHz.The first pulse in each record is concerned withthe direct path wave travelling in a lateral direc-tion from the echo-sounder source to the hy-



Figure 2. Experimental set-up. The combined echo-sounder/hydrophone moves along a straight line, receiving a signal everycentimetre. After being filtered and amplified, this signal was converted into a digital format to be processed. Compensa-tion for the beam-spreading (Time Variable Gain) and squared echo integration could thus be processed as the echo-

sounder moves

Table I. Features of the transducers used in the experiments

Transducer Frequency Bandwidth Source level Directivity Beam width PulsekHz kHz dB re 1Pa 1 m index dB (1/e) duration s

IA-100 102 96-108 204 12.3 29 50Raytheon V700 201 198-204 211 13.3 16 200Ulvertech 295 527 514-539 233 18.3 6 100


4/12

drophone. A 50-s pretrigger was set in the acqui-sition system.

The reflected echo produced by the sand/medi-um discontinuity surface reflection can be easilyseen at t 1 050 s. The echo signal shows anoth-er peak pulse that appears later (t 1 100 s), at100 and 500 kHz frequencies belonging to the bot-tom tank reflection. It is noteworthy that at 200 kHzfrequency this pulse is masked, since the length of

the incident pulse is longer than the sand layerthickness itself. The time arrival of these echoesdoes not vary significantly as the echo-soundermoves, because of the flatness of the bottom layerof sand.

A weak echo from the seaweed (700 s) can beseen just before the strong echo corresponding tothe sandy bottom (1 050 s). In this case the inci-dent pulse finds two targets (seaweed and thesandy layer), and both reflect or scatter acoustic en-ergy towards the hydrophone. At first glance, the

seaweed echo is not easily appreciable. Taking lay-er thickness into account, the arrival-time intervalfor seaweed echoes approaches that of sand (700-1 000 s). Their amplitudes, as can be observed,are much smaller than those from the sandy bot-tom. Usually at a frequency of 500 kHz, but alsosometimes at 100 kHz, the seaweed response canbe appreciated, but even then the noise and rever-beration sometimes mask this echo.

The 90 signal records were to create echogramplot. To improve visual detection of the seaweed,because of the weakness of its echo, a temporal win-dow (550-1 050 s) was processed.

The envelopes of all these signals have beencomputed with the Hilbert transformation (Carband Ranz, 1986). Graphs in figure 4 represent the90 envelopes of scattered signals in the region

where the seaweed echo arrives (570-1 070 s). Athin, dark layer appears in the last 50 ms, caused bythe sandy bottom reflection.



Figure 3. Sandy bottom back-scattered signal covered by a Gelidium seaweed layer using three different echo-sounders


5/12

The graph represents a M = 25 isolines (the lev-els have been normalised by the maximum ampli-tude of all the envelopes, m = 1, 2,... M). Such plotsdemonstrate stronger seaweed back-scattering at100 and 500 kHz.

To assess such visual appreciation, the followingacoustic parameters have been quantified: reflec-tion coefficient, absorption coefficient andacoustic impedance.

To detect any biomass present in the water col-umn, e.g. seaweed, with all the difficulties inherentin an ocean survey, including bad weather condi-tions, it is important to previously characterise thetarget in the laboratory very carefully. In order tocharacterise a target acoustically, it is necessary tomeasure at least two fundamental magnitudes: re-flection and absorption (expressed by the reflec-tion coefficient and the absorption coefficient, re-spectively). Another parameter of great interest isacoustic impedance, as it provides an idea of the

targets acoustic behaviour, when compared withwaters acoustic impedance.

Therefore, two methods are proposed to detectthe presence of the Gelidium biomass. First, bymeans of reflection. This method evaluates the en-ergy from the Gelidiumecho, in comparison withthe incident wave energy, to establish a certain ra-tio between both sound waves energies (Eg/Eo).

With this method, whenever the sonar of the ship

on a survey is passing over Gelidiumseaweed, wouldbe found same energy ratio, and then could be saidthat Gelidiumseaweed has been detected.

The second method is related to the absorptionof the wave energy that propagates through theGelidium seaweed layer, reflects on the sandy bot-tom, and again propagates through the Gelidiumseaweed towards the receiver. Comparing the ener-gy received at the hydrophone in two different cas-es, first passing through the Gelidiumseaweed, andthen when there is only sand (situation a) without



Figure 4. Echogram contour lines map of the back-scattered signal envelopes gated between 700 and 1 050 m


6/12

seaweed, the absorption produced by the Gelidiumlayer can be evaluated.

To evaluate both detection methods it is neces-sary, as noted above, to determine, through verycareful measurements, three main features of thetarget: reflection coefficient, absorption coeffi-

cient, and acoustic impedance.

Reflection coefficient

Marine acousticians often treat the integral ofthe squared acoustic pressure in a temporal win-dow as if the sound was reflected at the bottom.The process is called specular scatter, because it de-scribes scattering measurements in the specular di-rection. Using the image reflection equation as a

model (Clay and Medwin, 1977), the mean squarereflection coefficient (R2) is:

>R2< =R

(

02

pD

2

2dtp)

02z2

dt [1]

where p is the peak-to-peak amplitude of the echo,po is the peak to peak amplitude of the transmittedsignal at the distance R0, z is the layer depth and Dis the source directivity which is a measure of thedirectional properties of the transducer.

The measurements of the echo amplitude forscenario a, only sand without seaweed, obtained inthe laboratory experiments, were applied at theabove equation to compute the reflection coeffi-cient of the sandy bottom (Rs). Because of the sandybottoms echo arrival time, the temporal windowused in the aforesaid integral was (1 000-1 120 s).

The same process was applied to computing theaverage reflection coefficient of the Gelidium sea-

weed layer (Rg) with the algae in the water tank ly-ing above the sandy bottom (scenario b). The tem-

poral window corresponding to the seaweed echoarrival, ranges from 700 to 1 000 s.To compute the reflection coefficient of the

sandy bottom in the presence of Gelidium seaweed(Rs+g), scenario c, we estimated the ratio of thesand echo pressure amplitude in the presence ofGelidium, to that of sand without seaweed. Table IIshows the average of all the 90 values of the reflec-tion coefficients coming from the three differentsituations already described: a, b and c, and for thethree frequencies studied.

As can be seen in table II, the values of the sea-weed reflection coefficient are much smaller thanthe sandy bottom reflection coefficient; approxi-mately 10 times smaller. The influence of theGelidium seaweed layer it is not very important, ascan be seen comparing Rs+gwith Rs, meaning thatin this frequency range and with the thickness stud-ied, the seaweed layer is almost transparent tosound waves.

Acoustic impedance

The ratio between target impedance and waterimpedance gives an idea of how different, inacoustic terms, the target is compared with water. Ifthe ratio is small, it means that the behaviour of thetarget with regard to the incident sound wave doesnot differ greatly from the surrounding medium. It

would be hard to distinguish such a target from wa-

ter, acoustically.The opposite happens if the impedance ratio be-tween both target and water is large enough. Inthat case it would be possible, by reflective or ab-sorptive means, to detect the target submerged in

water.In scenario a, there is a discontinuity surface be-

tween the water and the sand. Using the reflectioncoefficient of such a discontinuity surface, the sandimpedance can be computed in relation to acoustic

water impedance.

Water impedance is defined by:Zw= wcw [2]

Analogically, the sandy bottom acoustic imped-ance becomes:

Zs = scs [3]

It is important to note that, instead of this, sandimpedance can be computed as a function of waterimpedance and the reflection coefficient. Thus,



Table II. Sand, G. sesquipedale, and sand with G. sesquipedalereflection coefficients

FrequencyReflection coeff. 100 kHz 200 kHz 500 kHz

Rs 0.677 0.429 0.402Rg 0.053 0.025 0.045

Rs+g 0.620 0.425 0.461


7/12

sand impedance is obtained with a non-intrusivemethod, very useful in cases where it would not beeasy to take samples from the bottom (deep wa-ters).

Z

Z

w

s

=

1

1

+

R

R

s

s

[4]

For scenarios b and c there are two different dis-continuity layers, one corresponding to the discon-tinuity between Gelidium seaweed and water, andthe other separating the sand and Gelidium sea-

weed. The reflection coefficient produced at thefirst discontinuity (seaweed/water), makes it possi-ble to compute the acoustic impedance of Gelidiumseaweed (Zg):

ZZ

w

g

=

11

+ R

R

g

g

[5]

Table III shows the average values of the acousticimpedance of the sand and the seaweed at thethree different frequencies used.

the reflection coefficients shown above. The ab-sorption coefficient can be expressed by theequation:

= 21d lnP

P

g

g

s (1 Rg

2)RRs+

g

g [6]

The values shown in table IV have been com-puted this way. In underwater acoustics, absorp-tion is usually expressed as dB/m, and Np/m isconverted to dB/m by multiplying (Np/m) by8.67.



Table III. Sand/G. sesquipedaleand water/G. sesquipedaleim-pedance ratios

FrequencyAc. impedance 100 kHz 200 kHz 500 kHz

Zs/Zw 4.242 2.665 2.356Zg/Zw 1.069 1.029 1.101

Table IV. G. sesquipedale, water and sand absorption coeffi-cients

Absorpt. FrequencyCoef. (dB/m) 100 kHz 200 kHz 500 kHz

Gelidium s. 1.60 1.22 3.23Water 0.031 0.059 0.125Sand 35 70 175

The fact that Zg/Zw is close to one reveals thesimilarity between water and seaweed imped-ance.

Absorption coefficient

In scenarios b and c, it has to be considered that

the echo produced by the seaweed/water disconti-nuity surface and received by the hydrophone haspassed twice (forwards and backwards) throughthe Gelidiumseaweed layer, suffering an absorptionexpressed by the quantity e2d, where is the ab-sorption coefficient (Np/m) of the Gelidium layer,and d is the thickness of the layer.

The absorption coefficient can be computedthrough a relationship (appendix II) between thepressure amplitudes of the sand echoes in sce-narios a and c (with and without seaweed), and

This table also includes values of the absorptioncoefficient for sand (Hamilton, 1987) and water(Fisher and Simmons, 1977) given by other au-thors, so that they can be compared with the onesobtained in the present paper.

DISCUSSION

The frequency range chosen is in good agree-ment with Clays biomass pyramid (Clay andMedwin, 1977). For marine plants and animals withdimensions between 2 cm and 20 cm, the effectivefrequency detection band varies from 25 kHz to250 kHz.

The findings of our laboratory experimentslead us to conclude that the reflection coefficientof Gelidium seaweed is very weak, approximately10 times smaller than the sand reflection coeffi-cient, and at the frequency range studied a sig-nificant dependence on frequency has not beenfound. This result is in accordance with theplankton echogram presented by MacLennan(MacLennan and Simmonds, 1992), where thediffuse background traces are often seen on theechogram, caused by the echoes from dense con-centrations of small creatures, e.g. plankton andmicronekton. Although the individual plankters


8/12

Figure 5. Emitted pulse, frequency response and directivity pattern of transducer IA-100

APPENDIX I

Transducers and hydrophone performances

are small or even microscopic, the overlappingechoes from a dense concentration produce sig-nals comparable to the ones given by individualfish.

We also found that there is not a major differ-ence between the energy reflected by sand when

there is Gelidiumcovering it, and when there is wa-ter only. The acoustic impedance of seaweed and

water are very similar. The acoustic absorption ofseaweed increases with frequency, being greaterthan that of water, but smaller than the acoustic ab-sorption of sand.

Once both methods proposed have beenanalysed, it can be concluded that seaweed detec-tion using sonar is quite difficult. First, because ofits very weak reflection coefficient, since the ratiobetween energies, reflected and incident, is so

small that it does not allow detection by means of

reflection. Second, because the absorption pro-duced by Gelidiumalgae is very low, making theseaweed layer an acoustically transparent biomass,so that it is difficult to detect it by the reduction ofthe sand echo energy whenever seaweed is pre-sent.

However, it should be pointed out here that,since the amount of power Gelidium diverts fromthe original wave varies with the layer thickness, theratio sand echo pressure/sand with Gelidium echo

pressure, formulated as PP

g

g

s e2d, should in-

crease whenever dincreases. With a sufficient thick-ness of the Gelidium seaweed layer, it could be pos-sible to detect a significant reduction of the energyreflected from the sandy bottom, and consequent-

ly, to detect the seaweed layer.



Amplitude(m

V)

Rela

tivelevel(dB)

Frequency (kHz)

Time (ms)

Directivity pattern


9/12



Figure 6. Emitted pulse, frequency response and directivity pattern of the Raytheon V-700 transducer

Amplitude(mV)

Relative

level(dB)

Frequency (kHz)

Time (ms)

Directivity pattern


10/12



Figure 7. Emitted pulse, frequency response and directivity pattern of the Ulvertech 295 transducer

Amplitud

e(mV)

Relativelevel(dB)

Frequency (kHz)

Time (ms)

Directivity pattern


11/12



Figure 8. Frequency response and directivity pattern of the Brel and Kjaer 8104 hydrophone

Bruel & Kjaer 8104 Hydrophone

Frequency Sensitivity at receptionkHz V per Pascal

100 30.2

200 3.8500 0.9


12/12



APPENDIX II

Absorption coefficient theoretical model

The transmitted pulse of sound propagates throughwater, encountering a few discontinuity surfaces (wa-ter/seaweed, seaweed/sand). These surfaces reflect ortransmit the pulse, and a portion of the incident ener-gy, proportional to the reflection/transmission coeffi-cient, returns to the hydrophone. The pressure ampli-tude of such a pulse, after been transmitted andreflected can be expressed by the equation:

Ps+g = TwgRs+gTgwe2dPo [7]

where Twg is the transmission coefficient of the wa-ter/seaweed discontinuity, Tgw is the transmissioncoefficient of the seaweed/water discontinuity,e2d express the absorption, and Po is the incident

pressure amplitude.The coefficient transmission multiplication is:

TwgTgw= 1 Rwg2 [8]

And the incident pressure amplitude is relatedwith the reflection coefficient by the expression:

Pg = RgPo [9]

Combining the three equations, an analytical ex-pression for the absorption coefficient can be ob-tained:

= (1/2d)ln

(Pg/P

gs) (1 R

g

2) (Rs+g

/Rg)

ACKNOWLEDGEMENTS

This research was financially supported by theComisin Interministerial de Ciencia y Tecnologa(CICYT) National R&D Plan, project PTR94-0158.The authors thank Dr J. Rey, E. Prez (Esgemar,

S.A.), Dr P. Siljestrm and A. Moreno (IRNASE,CSIC) for their assistance and helpful comments.

REFERENCES

lvarez, J., E. M. Llera, A. Vizcano, S. Gonzlez, I. Olivella,J. L. Catoira, X. M. Romaris, A. Borja, J. M. Salinas and S.Revenga. 1989. Programa de orientacin plurianual sobremacroalgas (1989-1991). Junta Asesora de CultivosMarinos (JACUMAR): 94 pp.

Carb Fit, R. and C. Ranz Guerra. 1986. Discrimination

temporelle dchos impulsifs qui se chevauchent.Traitement du Signal13: 145-151.

Clay, C. and H. Medwin. 1977. Acoustical Oceanography.John Wiley & Sons. New York.

Fisher, F. H. and V. P. Simmons. 1977. Sound absorption insea water.J. Acoust. Soc. Am. 62: 528-564.

Hamilton, E. 1987. Acoustic properties of sediments. In:Acoustic and Ocean Bottom.A. Lara, C. Ranz and R. Carb(eds.): 3-58. CSIC. Madrid, Spain.

Juanes, J. and A. Borja. 1991. Biological criteria for the ex-ploitation of the commercially important species ofGelidium in Spain. Hydrobiology 221: 45-54.

MacLennan, D. N. and E. J. Simmonds. 1992. FisheriesAcoustics. Chapman & Hall. London: 325 pp.

deteção de gelidium sonar

Documents