jfrc_1970_27_04.pdf (216.4Kb)

(1)

A

Comparison between

Two

Sonic Measuring Systems for Demessal Fish1

R. G. DOWD

Fisheries Research Board of Canada Marine Ecology Laboratory Bedford Institute, Dartmouth, N.S.

A N D

Institute of Marine Research Bergen, Norway

DOWD, R. G., E. BAKKEN, AND O. NAKKEN. 1970. A comparison between two sonic measuring systems for demersal fish. J. Fish. Res. Bd. Canada 27: 737-742.

Two sonic methods for estimation of abundance of fish stoclrs, the echo integrator and the digital counter methods, were compared on single and schooling fish in the Lofoten area of Norway during March 1969. Good correlation was obtained between the two systetns for both situations, but the slopes of the regressions of integrated values on the digital counter differed significantly between low and high density fish concentrations. This suggests that the two systems treated the echo information differently, but nevertheless maintained a linear relation between themselves over a wide range of counts.

Received January 23, 1970

INTRODUCTION

T w o MAIN PROBLEMS ARISE when acoustic instruments are used t o estimate fisl-i abundance. T h e first is t o obtain reliable counts of the number of fish per unit volume detected by the echo sounder. Different: systems have been developed for this purpose (Dragesund and Olsen, 1965; Dowd, 1967; Midttun and Nakken, 1968). T h e second is t o determine the effective volume of water sampled uniquely by each sonic pulse (Cushing, 1968). For the latter purpose, i t is necessary t o know the target strength of the fisli. Recently, Craig and Forbes (1969) have described a sonar for fis11 counting and derived a n excellent metliod for determining the number of fisl-i counted per unit volume.

T h e experiment reported l-iere is confined t o problems of the first kind and was carried o u t t o compare two kinds of signal processing systems, the Digital Echo Counter (Dowd, 1967) a n d a n Analog Echo Integrator (Dragesund and Olsen, 1965). T h e main questions t o be answered were: is there a linear

'~edford Institute Contribution No. 190.

737

Printed in Canada (J1699)

(2)

738 JOURNAL FISHERIES RESEARCH BOARD OF CANADA, VOL. 27. NO. ^4.1970

relation between systems when they are sampling signals froin the saiiic

"isonified" volume, and can these signals be correlated with a count of eclios obtained froin echo-sounder paper recordings? T h e two systems were COmpal-ed a t different pulse lengths with reflections from various densities of spawnirig cod (Gadus morhua L.). T h e work was done on board the Norwegian Researcll Vessel G. O. Sars, in Lofoten, Norway, in early Rilarch 1969.

EQUIPMENT AND METHOD

The equipineilt was set up as in Fig. 1. A Simrad Model EK38 scientific echo so~iiider was used with a transmitted power of 1000 w and a rneasured source level of

+

118 db/p b,lr a t 1 m, emitted from the 15 X 30-cm transducer. The receiver has a time varied gain (T.V.G.) function and was set to compensate two-way transmission loss down to 230 m. The outpiit o f the receiver was injected into the three recording systems: the echo-sounder recorder, the echo integrator, the digital echo counter. The threshold for all three was set a t -24 dbv to periiiit a direct coinparison between the systems.

EK 38 EK 38

RECEIVER RECORDER

F ~ ~ . 1. A sitnplified block diagram of the system iised on research vessel G. O. Sars. Arrows show

direction of signal in the system. INTEGRATOR

t-'--'

Fl

TRANSDUCER

DIGITAL COUNTER

The echo integrator (Dragesund and Olsen, 1965) accumulates the voltage from the received echos and resets automatically after each nautical mile or after full-scale deflectioii is reached. Also, it can be reset a t any time by the operator. The sainpling depth range is cotitrolled from the transmitted pulse, gating for the desired layer to be sampled. If individual fish ale recorded, the integrator reading is proportional to the number of recorded fish.

The echo-counting system (Dowd, 1967) records the number of echos from each tr'tii5- mission in digital forin. The system uses a delayed pulse technique to satnple a fixed distalite from bottom irrespective of water depth. The resolution obtained depends on the piilse lenqtli, and separate counts can be obtained for individual fish if they are separated by more thaii 25 cm. In this experiment the echo counter was changed to sample the same layer as the integratoi and converted from delayed piilse to transinit gate. The pulse lengths used were O 3, O 6, lii id 1 O msec, nominal. Since the integrator was not changed to compensate for change in piilie lengths and the pulse-length gates in the counting system were changed to O 6, O 9, aiid 1 3 msec respectively, a n increase in slope with an increase in pulse length was expected.

In Lofoten, the spawning cod are conce~ltrated in a layer between 75 and 125 in belotv the surface and where the depth of water is approximately 150 in. During the night the cod ( K -

curred as well-separated fish, but aggregated in shoals during the day. In a sample obtaiiied by purse seine the mean length was found to be 89 cni. The length of cod was geilerallv bet\+ecll 65 and 105 cm. The sampling of echo information was carried out mostly during the i l i ~ h t holli., and areas of different concentrations were selected. The ship speed was tnaintaiiled a t 5 l<llot\

The counter and integrator were adjusted to record signals from the same depth range, 100-125 111

The readouts were talren a t 10-inin intervals. Echo-sounder paper counts for the depth r,lIiw were taken whenever separation of echos permitted.

(3)

DOWD et al.: TWO SONIC MEASURING SYSTEMS

RESULTS

Figures 2A, B, and C sliow plots of integrator voltage versus echo count

1

with fitted regression lines. T h e correlation coefficients are 0.84 (24 df), 0.94 (30 df), and 0 . 9 6 (30 df), for the pulse lengtl-is of 0.3, 0 . 6 , and 1 . 0 msec respectively. Since the integrator voltage is directly proportional t o the pulse lengt11 and the counting system was adjusted t o coinpensate for tl-iis change, a n increase in slope with a n increase in pulse length was expected and obtained.

i

r) T h e slope for the three pulse lengths differed significantly a s follows:

Pulse length Difference between

(msec) Slope pulse lengths t d f

Figure 2D shows the relation between the integrator a n d counter for higli density daytime recording. T h e correlation coefficient between the two for , I the 0.3-msec pulse length was 0 . 9 9 (14 df). However, tl-iere was significant difference between the slopes for the day a n d night recordings (i.e., high and

l low densities) for 0.3-msec pulse lengths (t ⁼4.1, 37 df).

Figures 3A, B, and C sliow the relation between the actual paper counts and converted digital counts. Because individual counts cannot be resolved

I

on paper for high fish densities, the comparison was possible only for nigl-it recordings.

T o convert the digital instrument counts, C, into the estimated numbers of targets (fisll), N, we proceeded a s follows. If b, is the beain widtl-i a t the sampling deptli a n d t, is the transmission rate, if V is tlle ship's speed and T is the time of run, then the overlap factor between successive transmission

Uwt,

is (-). T h e estimated number of targets seen by the counting system

V T

during the sample period is then

With tllis formula, a conversion to the estimated numbers of fis11 was obtained and plotted against the paper counts (Fig. 3). Good correlation was obtained between the two methods with correlation coefficients of 0 . 9 6 (23 df), 0 . 9 8 (29 df), and 0.91 (30 df) for the

ul se

lengt11 of 0 . 3 , 0 . 6 , and 1 .O insec re-

l spectivel y.

DISCUSSION

T h e integrator readings and the echo counts for data of Fig. 2A-D could l have been converted t o the actual numbers of fish since the target strengt11 of cod is known. T11is is unnecessary wlien coinparing the two measuring

(4)

,740 JOURNAL FISHERIES RESEARCI-I BOARD OF CANADA. VOL. 27, NO. 4, 1970

:l.

-

.- 7 6

E .?

G ;

it -

' Z C?

i % n

- 3

. .2 d

-

⁵2 GT

5

2

-

, c j

' a,

"C,

M m

3 _a

- .

^

-

.. ^a

M .z E 2 ?.

r,

,-x

2%

L

2 c a

E w

c _c

3 a,

2 3

2 s

u .r:

:n

u i

.- tc,

sa,

-2

u

2 "j

eia'

trf:

z t'

i>

9 bo +.5 8'Li

2 g

-0 L

. 2

-

^{. V}

c

a, G

%!

4 2 E

- m E!

2 0

. d

+. -_

z C

-.

_uL-,

. + -1

v M

(5)

DOWD et al.: TWO SONIC MEASURING SYSTEMS

I ^c ⁸ l , , , , , , , , I , , , . , , , . 1

I 10 100 1 0 0 0

ECHO-SOUNOER PAPEii COUNT

' O O 0 l

C- NIGHT, 1.0 MSEC

,

¹⁰⁰ ^{I O W}

l

ECHO-SOUNOER PAPER COUNT

1.

FIG. 3. Relation between converted digital echo count and echo-soutlder recorder paper count for: (A) 0 . 3 msec, ( B ) 0.6 tnsec, and ( C ) 1 . O msec.

systems. A t low fis11 densities, good agreement was found in all instances among the three systems. However, the relation between the number of counts and t h e integrator values was found t o be significantly different between d a y and night. This implies t h a t one or botli of tlie recording systems are not linear with t h e actual numbers of fish. At night, when fis11 are recorded a s singles, bot11 units record values proportional t o the actual numbers. B u t b y day,

i

when t h e fis11 are aggregated, the digital recorder accumulates lower numbers in relation t o the integrator. However, the relation between the two systems

i

(6)

742 JOURNAL FISHERIES RESEARCH BOARD OF CANADA, VOL. 27, NO. ^4,1970

is linear over a wide range of total counts, implying t h a t the density of aggregation is reasonably constant and independent of school size. It als0 implies tl-iat neither system has a linear response t o changes in the total actual numbers of fish during schooling. Further experiliients should be conducted on the aggregation behaviour of fisli for stock estiination, including integration, counting, and fishing, t o attempt t o obtain a better Irnowledge of the re1 a t' ioii between sonic measurements and the actual numbers of fis11 is scl~ools.

ACKNOWLEDGMENTS

We thank Mr L. Midttun of the Institute of Marine Research, and the officers and crew of the Norwegian Research Vessel G. O. Sars for their assistance in inaking the experimeilt pob- sible. Also we thanlr Dr L. M. Dickie, Dr R. W. Trites, and Mr T. Platt whose criticism aiid assistance, during writing, were invaluable.

REFERENCES

CUSHING, D. H. 1968. The abundance of hake off South Africa. Fish. Invest. Min. Agr, Fish. Food (Gt. Brit.) Ser. 11. Salmon Freshwater Fish. 25(10): 1-20.

CRAIG, R. E., ^ANDS. FORBES. 1969. A sonar for fish coiinting. Fiskeri Dir. Skr. Ser. Hav- unders. 15: 210-219.

DOWD, R. G. 1967. An echo counting system for detnersal fishes. FAO Conference on fish behaviour in relation to fishing techniques and tactics. FB/67/E/7: 1-6.

DRAGESUND, O., AND S. OLSEN. 1965. On the possibility of estimating year-class streiigth b measiiring echo-abundance of O-group fish. Fiskeri Dir. Slrr. Ser. Havunders. 13(8):

47-75.

MIDTTUN, L., ^{A N D}O. NAICICEN. ,1968. Coullting of fish with an echo-integrator. Iiit. Couiic.

Explor. Sea B17: 1-7.