You are on page 1of 35

Reviews in Fish Biology and Fisheries 7, 134 (1997)

Underwater acoustics in marine fisheries and


fisheries research
O L E A RV E M I S U N D
Institute of Marine Research, PO Box 1870, N 5024 Bergen, Norway. E-mail: olem@imr.no

Contents
Abstract
Introduction
Application in commercial fisheries
The echosounder
The sonar
The trawl sonde and gear sensors
Fish abundance estimation
The echo integration method
Linearity
Instrument accuracy
Backscattering cross section
Influence of orientation
Swimbladder dependence
Size dependence
Frequency dependence
Effects of shoaling
Acoustic shadowing
Vessel avoidance
Long-range avoidance
Short-range avoidance
Variability of avoidance
Avoidance of vessel light
Allocation to fish species and size
Acoustic classification of fish species and size
Spatial variability
Density distribution
Replicability of survey estimates
Comparison with other independent methods
The echo counting method
The sonar method
Fish behaviour and distribution studies
Mapping of fish movements by acoustic tags
Recording of fish behaviour by echosounders and sonar

09603166

1997 Chapman & Hall

page 2
2
3

22

2
Summary and prospects
Acknowledgements
References

Misund
24
26
26

Abstract
Underwater acoustics enables the detection and precise location of fish and is therefore a
prerequisite for effective fishing methods such as pelagic trawling and purse seining. The
application of acoustic instruments to detect fish and monitor gear performance in
modern commercial fisheries is outlined. The latest developments in obtaining
information such as bottom roughness and determining such characteristics of fish
detected as size and species are presented.
Echo integration is now widely used to estimate the abundance of commercially
important fish stocks. The principles of the method are outlined briefly, and special
emphasis is put on such effects of fish behaviour as the dramatic influence of fish
orientation on its backscattering cross section, the possible effects of vessel avoidance,
and the uncertainties connected with spatial variability.
The use of acoustic tags, echosounders and sonar to study and quantify fish
behaviour and distribution is outlined, with particular attention to new developments
that provide detailed information on fish behaviour and distribution in relation to
environmental parameters.
Future developments and improvements in the application of underwater acoustics to
commercial fisheries and fisheries research are suggested.
Introduction
Some of the first-known trials for applying underwater acoustics to detect fish at sea
were conducted in the 1930s at the spawning grounds of the Barents Sea cod in Lofoten,
northern Norway (Sund, 1935), and during the driftnet fishery for herring off England
(Balls, 1948). Since then, location of fish by underwater acoustics has become an
increasingly important aspect of commercial fisheries. Much of the fish taken by trawlers
and purse seiners is located first by echosounders or sonars. In some fisheries, such as
the North Atlantic purse seine and pelagic fisheries and the Bering Sea pelagic trawl
fisheries, the capture is totally dependent on location of fish by acoustic instruments.
After the Second World War, underwater acoustics gradually became an important
method in fisheries science for mapping the distribution of commercially important fish
stocks (Devold, 1950), and the reflecting properties of fish began to be explored
(Midttun and Hoff, 1962; Shibata, 1971). The need for a more quantitative method led
to the invention of the echo integrator (Dragesund and Olsen, 1965), in which the
voltage generated by the backscattered echo signals is squared and summed over
intervals of depth and distance sailed. By calibrating the echo integrator unit using
metal spheres with known backscattering strength (Foote et al., 1987), the recording
properties of the instruments can be measured. If the backscattering strength of the
recorded fish is known, the echo integrator output can be converted to units of fish
density (MacLennan and Simmonds, 1992). These principles are now applied in regular
acoustic surveys for estimating abundance of economically important fish stocks
worldwide.

Underwater acoustics in marine fisheries

The implications for world fisheries of the development of underwater acoustics have
therefore been twofold. Acoustic fish detection equipment has enabled fishers to locate
and catch fish far beyond the sustainability of most commercially interesting fish stocks
in the world. On the other hand, by scientific application of acoustic equipment, the
distribution and abundance of fish stocks can be mapped and estimated independent of
commercial fishing activity. These scientific estimates are becoming increasingly
important in the management of many commercially important fish stocks.
In this review, I focus on the application (1) of acoustic instruments to locate fish
and monitor gear performance in commercial fisheries, and (2) of acoustic methods in
fisheries research to estimate the abundance of fish stocks and to record fish behaviour
and distribution. The physical principles and technology behind underwater acoustic
instruments and methods are only briefly covered. Interested readers will find a
complete coverage of these topics in Clay and Medwin (1977) and Urick (1983) on the
physical principles of underwater acoustics, in Mitson (1983a) on the principles and
technology of fisheries sonars, in Ona (1994) on the latest developments in fisheries
acoustic instrumentation, in MacLennan and Simmonds (1992) on acoustic methods in
fisheries research, and in Aglen (1994), Rose (1992) and Traynor et al. (1990) on the
sources of errors associated with application of underwater acoustics for measuring fish
abundance. The development of fisheries acoustics has also been considered in four
international symposia organized by the International Council for the Exploration of the

Sea (ICES) starting in Bergen, Norway, in 1973 (Rapports et Proces-Verbaux des

Reunions, 170), then in Bergen again in 1982 (FAO Fisheries Report No. 300; Rapports

et Proces-Verbaux des Reunions, 184), in Seattle, USA, in 1987 (Rapports et Proces


Verbaux des Reunions, 189), and finally in Aberdeen, Scotland, in 1995 (ICES Journal
of Marine Science, 53).
During the development of methods and instrumentation for underwater acoustics in
fisheries and fisheries research, there has been an increasing awareness of the effects of
fish behaviour, exemplified by concern about the dramatic influence of fish orientation
on its backscattering cross section (Foote, 1980a), the possible effects of vessel
avoidance (Olsen, 1979, 1990), and spatial variability (Aglen, 1983, Freon et al., 1993).

In this review, the effects of fish behaviour in fisheries acoustics are specially
emphasized.
Application in commercial fisheries
THE ECHOSOUNDER

Most fishing vessels, small or large, from all gear categories, are equipped with an
echosounder (Fig. 1). The sound beam is usually transmitted vertically, and the operating
frequency may vary from about 12 kHz up to about 200 kHz. The fishers use the
echosounder to record the depth to the bottom and to locate fish in the water column
below the vessel. Most modern echosounders have a colour display with colour indication
of echo strength. Strong echoes are displayed in red or brown, while weaker echoes are
presented in blues or greens. On older echosounders the recordings were indicated on a
black-and-white paper record; strong echoes were indicated in black while weaker echoes
were given a grey shading. Some modern echosounders still have the possibility to print
the echogram on paper black-and-white or colour. The great advantage of the paper
record is that it can be preserved for the future, and used to demonstrate or study

Misund

Fig. 1. Acoustic equipment for recording fish and sea bottom, and monitoring gear performance on a
modern pelagic trawler.

interesting observations. This is not always of interest during commercial fishing, where
the outcome depends largely upon quick tactical decisions based on recordings as they
are obtained. Echo recordings of such short time intervals are best provided through a
colour display on a monitor.
Some echosounders can operate at two frequencies simultaneously (usually a low one
around 50 kHz and a high one around 200 kHz). Experienced fishers use this feature to
distinguish between species, because the backscattering properties of some species are
frequency dependent. For example, in the North Sea, herring (Clupea harengus,
Clupeidae) and Atlantic mackerel (Scomber scombrus, Scombridae) may frequently
occur on the same fishing grounds, but fishers can distinguish their acoustic images
because herring shoals appear as strong (red) recordings both at high and at low
frequency while mackerel shoals give strong (red) echoes on high frequency but weak
(green) echoes on low frequency.
At present, there is a trend towards implementing software functions that may extract
more information from the received echo signals. Sophisticated echosounders may offer
a relative quantity of the recorded echoes, an approximate size distribution of those fish
recorded as single echoes, and may even reveal the direction of movement of single
fish as they move through the sound beam (Bodholdt et al., 1989). To increase the
volume searched, there are echosounders that transmit one beam slightly to either side
of the vessel in addition to the vertical one.
The application of echosounders in fisheries may also have functions other than
detecting fish and determining the depth to the sea bottom. During trawl and Danish
seine fishing, knowledge of the bottom type is of vital importance. An echo signal
processor that can be connected to commercial echosounders and that provides
information on the hardness of the bottom has therefore been developed (Burns et al.,
1989). This echo signal processor, named RoxAnn, displays the signal strength of the
first oblique bottom echo (from transducer to bottom to transducer) and the second
oblique bottom echo (transducerbottomsurfacebottomtransducer). A hard bottom

Underwater acoustics in marine fisheries

substrate like rocks and stones will produce a high ratio between the second and the
first bottom echoes, whereas a soft substrate like mud will produce a low ratio. Other
bottom substrates will produce ratios intermediate between those of mud and rock,
according to hardness. The RoxAnn echo signal processor has become a useful
instrument for fishers conducting trawling or Danish seining in waters with bottom
substrates where it is difficult to conduct these kinds of fishing methods, and during
scallop dredging. A few Norwegian scallop dredgers have installed an expensive
multibeam echo sounder (Simrad EM-12) that can map the countours of the sea bottom.
T H E S O NA R

Most purse seiners and pelagic trawlers fishing for schools and shoals (for distinction
between these terms see Pitcher, 1983) that are not visible on the surface are equipped
with a sonar that has a tiltable and trainable beam (Fig. 1). The sonar can operate
according to the searchlight principle in which a single beam is trained in sequential
sectors with one or more transmissions in each sector. More expensive instruments are
multibeam or omnidirectional, in which several beams cover a large sector or the whole
circle in each transmission. With such instruments, the complete horizontal extent of fish
shoals may be projected for each transmission (Fig. 2). Some sonars may also be
operated with an additional vertical beam fan which enables the vertical extent of fish
shoals to be projected as well.
Sonars constructed for detecting targets at long range operate at low frequency (from
around 20 kHz up to about 50 kHz), and have a beam width of around 108. This gives
rather limited resolution in the display of the recordings. High-frequency sonars, which
operate from about 150 kHz to about 200 kHz, give better resolution because they have
a narrower beam width of about 58. The accuracy with which a multibeam sonar may
project the dimension of an object across the beams will be in the interval
is the beam
[0, 2R tan ( )], where R is the horizontal distance to the target and
width of the sonar (Misund, 1994). Therefore, the narrower the beam, the better the
resolution in the projection of the targets. Purse seiners are often equipped with both a
low-frequency sonar for detection of fish shoals, and a high-frequency one to obtain a
better image of the extent of the shoal. The high-frequency sonar will also give an
image of the net wall of the purse seine, and so may be used to follow the movements
of fish shoals in relation to the net during pursing. However, heavy manoeuvring that
produces a flow of propeller water full of air-bubbles around the vessel will often block
school detection during this critical stage of the fishing operation.
The detection range of low-frequency, wide-beam sonars is mainly limited by
reverberation from various sources such as the surface, the bottom and from scattering
objects in the water column. Increased absorption reduces the detection range of highfrequency, narrow-beam sonars to at least half of the low-frequency ones. To enhance
the detection range, some sonars can employ frequency-modulated transmission. On the
Simrad SA950 and SR240 this functions as a transmission of a combination code of
two to eight different frequencies. Returning echoes are only displayed if they appear
with the right frequency code. This transmission principle results in a substantial
rejection of noise signals, and increases the detection range of weak targets.
When conducting active fishing such as purse seining, the use of sonar also has a
tactical purpose. After detection of fish, the sonar is used to keep contact with the
shoal to estimate its size, and to track its swimming depth, speed and direction of

Misund

Fig. 2. Recordings of herring schools by the Simrad SA950 sonar as reproduced by UNIRAS
software. The sonar beam is transmitted in a 458 sector and received by 32 adjacent single beams of
1.78 each (between 3 dB points). The orthogonal beam width is 108 ( 3 dB points). The recordings
are displayed with a colour representing a value from 0 to 63 (scale at left), and the number of the
colour value increases with higher fish density.

movement. During this process, the vessel is often manoeuvred in a circle around the
shoal. To give the fishers the best possible information about the capture situation,
many sonars have such sophisticated functions as automatic target tracking and display
of operation parameters such as distance, bearing, depth, speed and heading of the fish
school (Bodholdt, 1982). Several sonars can also display a true-motion picture which
presents the movements of both the fish school and the vessel (Bodholdt and Olsen,
1977). Models exist, and can be implemented in the software of sonars, for predicting
the swimming behaviour of fish schools in purse seine capture situations (Misund,
1994).
Sonar-guided fishing is now conducted by purse seiners when fishing small and
medium-sized pelagic species of the families Carangidae, Clupeidae, Engraulidae,

Underwater acoustics in marine fisheries

Scombridae and Salmonidae worldwide. Still, there is probably potential for extending
the method when using purse seines to catch other species. Tuna purse seining is
primarily based on visual detection of surface schools. Strong binoculars, radars
detecting birds that associate with fish aggregations on the surface, helicopters,
airplanes and airborne radar systems (Anon., 1992) may be used to enhance the
detection range. In the eastern Pacific, the fishery for surface schools is based on
detection of dolphins, with which tuna associate (Anon., 1992). The capture tactic is
based on encircling both dolphins and tuna by the purse seine, and subsequent release
of the dolphins by the backdown procedure. In this process, dolphins may occasionally
become entangled in the net and drown. The total annual mortality of dolphins in this
fishery was about 60 000 individuals in 1990 (Anon., 1992). Fishers in the eastern
Pacific avoid using sonar because the tuna tend to follow the dolphins, which can
detect frequencies up to about 150 kHz (Au, 1993) and which therefore hear the sonar
at long range and move away from it. However, it is possible that sonar could be used
to detect and guide the capture of tuna schools that are not associated with dolphins.
Such sonar-guided purse seining for tuna is conducted with success in the western
Pacific (Anon., 1992), and it is possible that the method might have contributed to a
reduced dolphin mortality in the fishery in the eastern Pacific.
In areas with a temperature-stratified water column, the detection range of sonars
may be severely limited. A common situation in temperate regions during summertime
and in the tropics, is that the sea temperature declines with depth. This causes a sound
beam that is tilted slightly with respect to the horizontal to bend downwards, because
the speed of propagation of the wavefront is reduced at colder temperatures (Mitson,
1983a). The result may be a substantial reduction in detection range of targets close to
the surface. In summertime, the maximum detection range of herring and mackerel
schools in certain areas of the North Sea may therefore be just about 400 m, while
schools may be detected up to 2000 m away in other areas with no temperature
stratification. Another source of variation is internal waves, which may locally introduce
a five-fold difference in detection range (Smith, 1977).
T H E T R AW L S O N D E A N D G E A R S E N S O R S

During the 1960s, pelagic trawling was developed as a capture technique for shoaling
species. To monitor the vertical opening of the trawl, and the vertical position of the
trawl in relation to fish concentrations and the bottom, the trawl sonde was invented
(Fig. 1). The transducer of this instrument is mounted on the headline of the trawl and
operated through a cable to the ship. On most instruments, the transducer can sound both
vertically and horizontally. In recent years, there have been developed alternatives to the
traditional trawl sonde. A cable-connected sector-scanning sonar operating on 330 kHz
with a beam of either of 28 3 308, or 28 conical, gives a detailed image of the trawl
opening and how it is positioned in relation to fish concentrations and the sea bottom
(Ona and Eger, 1989). For those who find the cable inconvenient, a cableless trawl sonde
with a vertically and horizontally sounding transducer mounted on the headline is
available.
To be able to follow how the gear is functioning during fishing, acoustic gear
performance sensors have been introduced during the last two decades. On trawls there
are now cableless acoustic sensors to measure trawl depth, door spread, headline height,
sea temperature, speed through the water and amount of catch in the bag. There are

Misund

also special depth sensors to follow the performance of purse seines during sinking and
pursing.

Fish abundance estimation


T H E E C H O I N T E G R AT I O N M E T H O D

The sound intensity (I0 ) of a pulse emitted from an underwater transducer decreases due
to geometrical spreading and absorption as it propagates through the sea. If a fraction of
the pulse is backscattered by a number of objects (N) per unit volume at a certain range
(R) in the sea, the sound intensity (I) arriving at the transducer is expressed by the sonar
equation, expressed as (Forbes and Nakken, 1972):
2

.e

2 R

b2 (, ) d

.R

I 0 . (c 2) .

.(

4 ). N

(1)

where: c represents the speed of sound underwater ( 1500 m s 1 ); is the duration of


the sound pulse ( 0.001 s); b is the beam function; is the beam width alongship; is
the beam width atwarthship; d is the solid angle; is an attenuation coefficient; and
is the backscattering cross section of the objects (m2 ).
To compensate for the transmission loss, the returning intensity is magnified by a
time-varied gain function equal to R2 . e2 R . The transmission loss due to geometrical
spreading occurs in both directions, so it should ideally have been compensated by an
amount equal to R4 , and this is done during target strength measurements of single fish
or of solid spheres during calibration of the echo integration unit. When a sound pulse
emitted from the transducer propagates through the sea, the area of the pulse increases
with range (R) at an amount equal to R2 . The larger the area of the pulse, the greater
the number of fish in a layer with constant density that may be insonified. To correct
for this range dependence in the number of fish that may be insonified, the
compensation for the transmission loss is reduced from a factor proportional to R4 to a
factor proportional to R2 .
In an echo integrator the voltage (V) of the returning signals is squared and summed
up over defined range intervals to give an output (M) according to the equation (Forbes
and Nakken, 1972):

(R2 . e2

. V 2 ) dR

R1

R2

R2

(R2 . e2

. k . g . I ) dR

(2)

R1

where k denotes the receiving voltage response and g is the total gain of the echo
integration system.
If the echo integration unit is calibrated by the use of a metal sphere with known
backscattering properties (Foote et al., 1987), the echo integrator output can be
converted to fish density per unit area ( A ) by the equation (Dalen and Nakken, 1983):
(CI ) . M sA
(3)
A

where: CI is the calibration constant of the equipment; M is the echo integration output
(mm per nautical mile); and sA is the area backscattering coefficient (m2 /(nautical
mile)2 ).
During surveys, the echo integrator is set to produce outputs for predetermined depth

Underwater acoustics in marine fisheries

channels over a certain distance sailed (e.g. 0.1, 1 or 5 nautical miles). The area
backscattering coefficients (sA values) have to be partitioned to actual species according
to fishing samples (see Allocation to fish species and size, below) or other means of
identifying the recordings. The number (Ni ) of fish of a species (i) in an area (A) is
then found by:
(4)
Ni
Ai . A
:

The echo integration functions through continuous echosounding within a vertically


directed narrow cone along the survey transect (Fig. 3). Fish close to the surface (in the
upper dead zone) or close to the bottom (in the bottom dead zone) are not recorded.
The extent of the upper dead zone is about 10 m because the echo integration starts a
few metres below the transducer, which is usually mounted on the hull at about 5 m
depth (Aglen, 1994). The definite bottom dead zone is limited to within half a pulse
length from the bottom (Mitson, 1983b), and the acoustic dead zone is therefore a
function of bottom depth and beam width (Ona and Mitson, 1996). The echo
integration dead zone is then a function of the acoustic dead zone given by the bottom
detection, a backstep zone set immediately above the sea bed, and a partial integration
zone where only a portion of echoes from fish near the sea bed are detected (Ona and
Mitson, 1996). When surveying fish such as orange roughy (Hoplostethus atlanticus,
Trachichthyidae) distributed near step sea-mountains at about 1000 m depth, the extent
of the bottom dead zone will be especially significant (Kloser, 1996). Towing the
transducer in a body that can be lowered to 500650 m depth may reduce the extent of
the bottom dead zone by about 50% and thereby provide more precise abundance
estimates (Kloser, 1996).
The quality of recordings from hull-mounted transducers is weather dependent
because wind-induced air-bubbles may attenuate and even block the echosounder
transmissions (Dalen and Lvik, 1981; Novarini and Bruno, 1982). Mounting the

Fig. 3. Vertical recording windows for sonar, echo integration and bottom trawl during fish abundance
estimation surveys.

10

Misund

transducer in a towed body or on a protrusible keel that may extend 34 m down from
the hull may enable acoustic surveying even in rough weather (Ona, 1994).
A significant improvement in the echo integration method has been achieved by
computer-based post-processors (Dawson and Brooks, 1989; Knudsen, 1990; Weill et
al., 1993). With these instruments it is possible to replay the acoustic recordings, do
thresholding to separate scatterers, allocate area backscattering strengths to species,
identify schools, and correct the recordings for noise and bottom spikes.
Linearity
The echo integration method has been experimentally proven to be linear (Foote, 1983),
which means that the acoustically measured density and the real density are equal. This
principle was found to be valid for measurements of herring about 27 cm long in a
0.7 m3 cage at densities up to about 60 fish m 3 (Foote, 1983). Free-swimming herring of
this size usually organize themselves into schools at an average density of about 2
fish m 3 (Misund, 1993). Rttingen (1976) observed a linear relationship between echo
intensity and fish number up to a density of about 100 individuals m 3 for about 35 cm
long saithe (Pollachius virens, Gadidae) and about 2000 individuals m 3 for about 10 cm
long sprat (Sprattus sprattus, Clupeidae) held in a net cage of about 2.7 m3 . For higher
densities, the relationship became non-linear, and the coefficient of reflectivity decreased.
For the saithe measurements, Foote (1978) was able to deduce the mean extinction cross
section (backscattering cross section absorption cross section) of the fish. The
observed relationship between echo intensity and fish density could thereby be
reproduced fairly accurately as a function of the extinction cross section (Foote, 1978).

Instrument accuracy
By incorporating the general progress in electronics and computer technology, the
performance and accuracy of scientific echosounders and integration units have been
improved. Simmonds (1990) concluded that the accuracy of a 38 kHz Simrad EK400
echosounder varied within about 7% over a 5 year period. More modern echosounder and
integration units with a higher degree of computerization probably have even better
performance and accuracy.
Backscattering cross section
To convert the echo integrator output over a specific depth interval and distance sailed, it
is necessary to know the backscattering cross section ( ) of the fish. In the classic
literature of underwater acoustics this quantity is usually expressed as the target strength
(TS) of the fish, and is defined (Urick, 1983) as TS 10 log10 ( /4 ). With
contemporary technology of digitized echo integrators, it is often felt confusing to
sustain calculations in logarithmic units. The target strength of fish can be measured
experimentally on tethered, stunned individuals (Nakken and Olsen, 1977), on live fish in
cages (Edwards and Armstrong, 1983), or in situ by the dual-beam or the split-beam
technique (Ehrenberg, 1983). During in situ measurements, the angle from the acoustic
axis to the target must be estimated so that the beam pattern effect can be compensated
for. In the dual-beam method, this is done by applying one narrow and one wide beam so
that the angle to the target can be estimated on the basis of the relative backscattered
echo intensity achieved by the two beams. A split-beam transducer is divided into four
quadrants during reception mode so that the angle to single targets can be estimated by

Underwater acoustics in marine fisheries

11

use of the phase difference of the backscattered echo intensity on the four quadrants. It is
now recognized that the algorithms used for identifying single echoes in split-beam
echosounders have not been good enough, and have accepted multiple echoes from highdensity assemblages so that in situ target strengths obtained by such instruments have
been overestimated (Soule et al., 1995). Possible improvements to these algorithms have
been suggested (Soule et al., 1995). There also exist theoretical models for calculating
the backscattered cross section of fish as functions of swimbladder presence, fish size,
orientation and frequency (Foote, 1985; Furusawa, 1988).
Influence of orientation. Unfortunately, the backscattering cross section is very dependent
upon the incident angle between the sound pulse and the fish. The sound pulse from the
echosounder is transmitted vertically, and insonifies the fish in the dorsal aspect. The
fish has its maximum backscattering cross section when it is orientated nearly
horizontally, but a change in vertical orientation with the body tilted slightly upwards
or downwards may easily result in a reduction of the backscattering cross section by a
factor of 100 (Nakken and Olsen, 1977; Foote, 1980a), in extreme cases by a factor of
10 000 (Foote and Nakken, 1978). Changes in the roll angle of the fish also affect the
backscattering cross section, but not to such a dramatic extent as do changes in tilt angle
(Love, 1977).
Swimbladder dependence. The air-filled swim bladder accounts for about 9095% of the
backscattering cross section of fish (Foote, 1980b). Fishes that lack this organ, such as
the Atlantic mackerel, consequently have a backscattering cross section that is only about
1/10 that of comparable swimbladdered species (Foote, 1980b). The swim bladder can be
open (physostomous) or closed (physoclist). Physoclist species like the gadoids have a
rete mirabile, which is an organ for secretion/resorption of gas from the blood to the
closed swim bladder. For such species the main function of the swim bladder is supposed
to be regulation of buoyancy by maintaining a constant swimbladder volume (Blaxter and
Batty, 1990), but it may also serve to enhance hearing (Hawkins, 1993). However, the
buoyancy-regulating process is slow, and physostomous fishes are not supposed to be
neutrally buoyant during their often substantial diel vertical migrations (Blaxter and
Tytler, 1978). Such migrations may affect the volume and thereby also the backscattering
cross section of the fish. Nevertheless, Harden Jones and Scholes (1981) estimated that
the effect of varying swimbladder volume and thereby backscattering cross section
through vertical migrations of physoclist fishes was probably less than that of tilt-angleinduced variations in the backscattering cross section.
Physostomous species such as the clupeids lack a gas-producing organ, while the
salmonids have a small gas-producing organ. Most physostomous species have a
posterior canal from the swim bladder to the oesophagus, and an anterior canal from
the swimbladder to the anus (Blaxter and Batty, 1990). It is therefore supposed that the
swim bladder is filled by swallowing air at the surface (Brawn, 1962), even to such an
extent that an above-atmospheric pressure is built up in the swimbladder (the big gulp
hypothesis, Thorne and Thomas, 1990), but opening of physoclist swimbladders of fish
caught at shallow depths often indicates a substantial gas pressure in the swim bladder
that is difficult to explain just by swallowing. Nevertheless, physostomous fishes are
often characterized by substantial diel vertical migrations which undoubtedly may affect
the back scattering cross section. Ona (1990) found that the swimbladder volume of

12

Misund

herring followed a depth relation according to Boyles law, and Halldorsson and
Reynisson (1983) and Olsen and Ahlquist (1989) found a certain reduction in
backscattering cross section of herring with increasing depth. For physoclist species it is
supposed that the gas-filled swimbladder may have only a limited function to obtain
neutral bouyancy, functioning rather to enhance hearing (Hawkins, 1993).
Size dependence. The backscattering cross section increases by approximately the square
of the fish length (L). Size-dependent equations for the backscattering cross section
expressed through the target strength (TS) of the fish are therefore often of the form
TS 20 log L b. Based on regression of results from several in situ measurements,
Foote (1987) recommends using TS 20 log L 67.5 for the physoclist gadoids and
TS 20 log L 71.9 for the physostomous clupeids when conducting acoustic surveys
with 38 kHz echosounders. Measurements of the backscattering cross section for
clupeids, gadoids, other fish species, plankton and squid are summarized by MacLennan
and Simmonds (1992).
The directivity of the backscattering of the incident sound becomes narrower as fish
size increases (Foote and Nakken, 1978). This means that the backscattering cross
section becomes more sensitive to tilt angle variations as fish size increases.

Frequency dependence. Most measurements of backscattering cross section of fish are


made at 38 or 120 kHz. MacCartney and Stubbs (1971) found a certain increase in the
backscattering cross section with increasing frequency, but an influence of frequency
seem species dependent. Both Edwards et al. (1984) and Degnbol et al. (1985) found a
lower backscattering cross section for herring at 120 kHz than at 38 kHz. For mackerel,
Edwards et al. (1984) found a 50% higher backscattering cross section at 120 kHz
compared with that at 38 kHz. Generally, at higher frequency the tilt angle dependence
becomes more variable (Nakken and Olsen, 1977).
Effects of shoaling. The tilt angle distribution is supposed to become narrower and the
mean tilt angle closer to zero when fish are schooling compared with when they are
shoaling, and the result is probably a higher average target strength for schooling fish
than for shoaling fish (Foote, 1981). As a result, a diurnal variation in target strength is
expected, and has been observed for caged cod (Gadus morhua, Gadidae), herring and
mackerel (Edwards and Armstrong, 1983; MacLennan et al., 1989, 1990; MacLennan,
1990). The variation was especially large (about 5 dB) for mackerel because this
swimbladderless species tilts upwards at low swimming speed when shoaling at nighttime to keep an upward hydrodynamic lift (He and Wardle, 1986).
According to these findings, different target strength values should be applied during
acoustic surveys for fish that are schooling during the day and shoaling at night. This is
not yet practised because the validity of applying cage measurements to wild fish is
questioned (MacLennan et al., 1990). However, there are several indications of diurnal
changes in the social aggregation pattern of fish that affect their orientation and thereby
target strength. Tilt angle measurements of herring in a large net cage (Beltestad, 1974)
and when free swimming (Buerkle, 1983) showed a change from slightly negative tilt
during daytime to slightly positive tilt during night-time. For walleye pollock (Theragra
chalcogramma, Gadidae), the target strength has been observed to be 3 dB lower by
night than by day (Traynor and Williamson, 1983).

Underwater acoustics in marine fisheries

13

Due to the difficulties in measuring target strength in dense aggregations in situ, the
information necessary to differentiate target strength according to the different social
aggregation patterns is at present lacking (Foote, 1987). One possibility to overcome the
problem for some species is to use a special tag to measure the tilt angle of fish when
passed over by a survey vessel both when schooling and when shoaling (Mitson and
Holliday, 1990). The average target strength can be calculated according to the
swimbladder morphometry of the species and the tilt angle distribution observed (Foote,
1985). This may also be done directly from the tilt angle distribution if the scattering
properties of the species are known (Foote, 1980a).
Acoustic shadowing. In natural fish aggregations that have a high density and a large
vertical extent, substantial extinction of the sound energy emitted from an acoustic
transducer can occur (Lytle and Maxwell, 1983). This is often observed during surveys
by a weakening, and in some cases disappearance, of the bottom echo under large dense
schools. For acoustic stock assessment the consequence of such extinction is
underestimation of the fish density. Appenzeller and Leggett (1992) suggest that
acoustic shadowing when fish occur in dense schools may cause an underestimation of
biomass by up to 50%. As an additional theorem to the linearity principle, Foote (1983)
suggested an equation involving knowledge of the fish density and extinction cross
section to compensate for the extinction. Similarly, Lytle and Maxwell (1983) proposed
three approximations to correct estimates of high density, all involving knowledge of the
extinction cross section. By recording the attenuation in the bottom echo as a function of
the vertical extent of herring aggregations in a Norwegian fjord, Toresen (1991) was able
to fit equations to correct the fish density estimates for extinction. Foote (1990) has
extended his extinction-correction solution for application to aggregations which vary in
density both horizontally and vertically, and Foote et al. (1992) have developed a method
to determine the extinction cross section and thereby to correct for acoustic shadowing
when fish are distributed in dense layers over a flat bottom.
Vessel avoidance
According to the avoidance behaviour model of Olsen et al. (1983a), fish will react to
noise stimuli by increasing their speed and swimming radially with a downward
component away from an approaching vessel. Such swimming behaviour may result in a
horizontal dilution of fish density in front of the vessel.
Long-range avoidance. Echosounder observations from an independent small vessel
showed that polar cod (Boreogadus saida, Gadidae) in the Barents Sea avoided at about
150 m in front of an approaching survey vessel (Olsen et al., 1983b). It is possible,
however, that this species is very sensitive to vessel sound, because numerous recordings
on cod and haddock (Melanogrammus aeglefinus, Gadidae) in the Barents Sea (Olsen et
al., 1983b; Ona and God, 1990) indicate little or no avoidance more than 50 m in front
of research vessels running at normal survey speed. Similarly, there was no change in the
distribution of shoaling herring at some distance in front of the approaching survey
vessel during experiments in Norwegian fjords (Olsen, 1979, 1990; Ona and Toresen,
1988b), and no detectable change in fish concentrations in front of survey vessels during
studies on small clupeids in the Caribbean Sea which employed echosounders operated
from an independent vessel (Freon et al., 1990; Gerlotto and Freon, 1990, 1992).

14

Misund

Echosounder recordings of fish concentrations do not reveal slight changes in


swimming behaviour of individual fish as a survey vessel approaches. Recordings made
by omnidirectional or multibeam sonars indicate that horizontal dilution of the density
of schooling fish may occur more than 50 m in front of survey vessels, depending on
species and geographic area. Goncharov et al. (1989) observed that schools of jack
mackerel (Trachurus symmetricus, Carangidae) increased the swimming speed
remarkably linearly within an interval of about 350 to 50 m in front of a large
approaching vessel. This pattern was observed in three different sea areas. There were
clear differences in the reaction distance among the three areas, but the slope of the
increase in swimming speed with decreasing distance from vessel to school was much
the same. Goncharov et al. (1989) argued that differences in the sea temperature caused
the variations in reaction distance among the three areas. Another possibility is that the
conditions for sound propagation varied so that the fish became aware of the vessel at
different distances in the three areas.
By use of an omnidirectional sonar, Diner and Masse (1987) observed that schools of
herring in the English Channel, and sardine (Sardina pilchardus, Clupeidae) and
anchovy (Engraulis ringens, Engraulidae) in the Bay of Biscay tend to avoid
horizontally at about 150 m in front of an approaching survey vessel. However, most
schools seemed not to avoid if they were not in the path of the vessel. Nevertheless,
Soria et al. (1996) interpret an overrepresentation of schools detected 4070 m to the
side of the vessel by sonar to have been caused by lateral avoidance in front of the
vessel. Misund and Aglen (1992) observed that schools of herring and sprat in the
North Sea tend to be guided and swim in the same direction as the approaching survey
vessel. This herding effect is suggested to be due to the noise directivity pattern of the
vessel. Because of shadowing by the hull, there will be lobes of maximum intensity to
each side of the vessel (Urick, 1983; Misund et al., 1996). Just in front of the vessel
there will be a minimum. Fish that sense and react to the increasing sound intensity in
front of an approaching vessel will thereby be guided in front of it (Misund et al.,
1996). This is apparent from the distribution of the swimming angle of approached
herring and sprat schools relative to the vessel (Misund and Aglen, 1992). However,
there seemed to be no relation between swimming speed, avoidance pattern and
distance between vessel and school in the interval from about 1000 m to 50 m in front
of the vessel (Misund and Aglen, 1992; Misund et al., 1996).
Short-range avoidance. Just before, and as, the vessel passes over, the avoidance model
(Olsen et al., 1983a) predicts sudden escape reactions. This sudden avoidance can affect
acoustic stock assessment, both through a horizontal dilution of fish density in the path
of the acoustic beam, and by a lower target strength than predicted because the fish are
tilted downwards (Nakken and Olsen, 1977). The strength of the reactions depends upon
the depth of the fish, the size of the vessel and the vessel speed. Echosounder recordings
from small independent vessels have shown that this is the case for shoaling herring in
Norwegian fjords, and for cod, polar cod, and capelin (Mallotus villosus, Osmeridae) in
the Barents Sea (Olsen, 1979, 1990; Olsen et al., 1983b). The resulting reductions in
density underneath the survey vessels have been quantified to vary from 40% up to 90%.
In accordance with the avoidance behaviour model, the reactions seemed very dependent
on the swimming depth of the fish. That the avoidance reactions depend upon the vessel
size (and engine power) was deduced from comparison of reactions of herring to the 300

Underwater acoustics in marine fisheries

15

GRT R/V Johan Ruud and to the 900 GRT R/V G.O. Sars (Olsen et al., 1983a). How the
speed of the vessel may influence the avoidance reactions has been studied by Doppler
shift measurements of capelin when approached by a survey vessel (Olsen et al., 1983b).
The vertical swimming speed increased more or less proportionally with increasing
vessel speed.
Other investigations on cod, haddock and herring in the same areas revealed no or
just weak avoidance reactions to fisheries research vessels running at normal survey
speeds. By echosounder recordings from a small observation vessel, Ona and God
(1990) detected no avoidance reactions of cod and haddock that were approached and
passed over by a survey vessel. When recording herring concentrations by a vertically
scanning sonar mounted at various positions on the hull, Ona and Toresen (1988a)
detected no changes in the herring distributions when the sonar transducer was mounted
on the bow and just in front of the echosounder transducers. This suggests that the
herring performed no avoidance reactions that significantly affected biomass estimation
by conventional acoustics.
Similar studies on clupeid species in shallow-water tropical areas showed no
reduction in fish density during passage of small survey vessels either when sailing or
motor driven (Freon et al., 1990; Gerlotto and Freon, 1990, 1992). Schools of

Sardinella aurita, Clupeidae off Venezuela avoided vertically, but only in the upper
20 m when overpassed by a medium-sized survey vessel. This resulted in little
underestimation of the density as the average tilt angle during the dive was estimated to
be less than 108 (Gerlotto and Freon, 1992).

The significance of avoidance studies using an echosounder from a small observation


vessel would have been easier to interpret if recordings from the survey vessel during
the experiments were also shown and compared with the recordings from the
independent vessel. In future experiments, this should be done by using identical,
calibrated echosounders on both vessels, and by applying small echo integration
intervals horizontally and vertically so that eventual avoidance reactions can be
quantified.
Recordings by omnidirectional and multibeam sonars show that both horizontal and
vertical avoidance of schools can occur close to the vessel. Soria et al. (1996) found
that schools of anchovy and sardine in the Mediterranean avoided laterally or
compressed vertically when close to an approaching vessel, and fewer schools than
expected were recorded by the echosounder. Similarly, about 15% of the herring schools
in the North Sea that occurred in the path of the survey vessel avoided horizontally
when closer than 50 m in front of the vessel and were not recorded by the echosounder
(Misund and Aglen, 1992). The schools that avoided were smaller and were swimming
shallower than those recorded by the echosounder. Depending on the sea area, 3565%
of the jack mackerel schools observed by Goncharov et al. (1989) avoided horizontally
and dived about 30 m when approached by a large vessel. Diving of herring schools has
also been observed when they are overpassed by purse seiners (Misund, 1994).
However, all schools of spawning migrating capelin that occurred in the path of the
survey vessel were recorded by the echosounder (Misund, 1994).
Variability of avoidance. As vessel avoidance is claimed to be the outcome of vessel
sound stimuli and natural stimuli acting on the fish (Balchen, 1984), substantial variation
may be expected. Several biotic and physical factors may influence the response

16

Misund

threshold and response strength to vessel sound stimuli. Neproshin (1979) observed that
the distance at which Pacific mackerel (Pneumatophorus japonicus) reacted to
approaching trawlers varied substantially with time of day. The reaction distance was
less in the morning and the evening than in the middle of the day. This was explained by
arguing that jack mackerel avoids most when well fed and swimming in schools in the
middle of the day, and less when shoaling and feeding in the morning and evening. When
Pacific mackerel occurred in low-temperature water, the reaction distance was less than
when the fish were swimming in warmer water.
The avoidance behaviour may also depend on the life history stage of the fish.
Herring have been observed to react more strongly to an approaching vessel when on
spawning migration than when hibernating, feeding or on feeding migration (Misund,
1994). Similarly, herring have been observed to react less to approaching vessels the
better the sonar conditions and thus the distance at which the fish can detect the noise
from the vessel (Misund and Aglen, 1992).
Avoidance of vessel light
At night, avoidance can occur as a response to visual stimulation by vessel lights. This is
well known by fishers, and purse seiners fishing on dense shoals are very careful with
use of light during circling, shooting and pursing of the seine. In most cases only the
necessary regulation lights are used. The consequences for acoustic stock assessment of
avoidance reactions due to the light of survey vessels may be the same as for avoidance
reactions to vessel noise. Levenez et al. (1990) observed significant vertical avoidance of

tropical clupeids off Venezuela during new moon when running a survey vessel with a
500 W light on the bridge. With regulation lights only, the fish were distributed more or
less evenly from surface to bottom at about 50 m. When the light was turned on, the fish
descended and concentrated in a narrow depth interval close to bottom. This distribution
pattern could be generated by switching the light on and off. However, the mean echo
integral was the same with the lamp on or off. When conducting a similar experiment in
the same area, Gerlotto et al. (1990) observed no vertical avoidance, but found
indications of species- or size-dependent horizontal avoidance when the lamp was on.
The last experiment was claimed to be conducted during dark nights, but it was still
during a full moon period. This might have had an influence on the fish so that they
reacted less to light than during other moon phases. Even if these few observations
indicate minor effects on acoustic stock assessment due to avoidance reactions to vessel
light, survey vessels should be operated as dark as possible at night-time to minimize the
influence of such reactions. It is the experience of Icelandic survey vessels that excess
lights from the decks and cabins can cause such strong avoidance reactions of herring
that the acoustic measurements may be significantly underestimated (Halldorsson and
Reynisson, 1983).
Allocation to fish species and size
A basic assumption of fisheries acoustics is that samples by trawl are representative of
the populations recorded. This principle thereby enables an unbiased allocation of
integrated echo intensity to different species and size groups (Dalen and Nakken, 1983).
Accordingly, fish are assumed not to react on the approach of the sampling gear in a way
that alters the species and size distribution present.
Vessel avoidance reactions during sampling may affect acoustic stock estimation

Underwater acoustics in marine fisheries

17

indirectly by altering the distribution of species and size groups available to the gear.
Similarly, such reactions may cause difficulties in combining the results from sweptarea and echo integration estimates of semipelagic fish (Nunnallee, 1991). When
exposed to playback of the noise of an approaching trawler, penned cod polarized and

dived slowly, but the reactions varied substantially (Engas et al., 1995). Ona and God
(1990) observed that the avoidance reactions of cod and haddock were stronger when
the survey vessel was towing a trawl, and that diving made the fish more available to
bottom trawl and less available to pelagic trawl. For Pacific whiting (Merluccius
productus, Gadidae), Nunnallee (1991) observed no change in vertical distribution, but
a higher echo value probably due to a change in orientation of the fish when a trawler
passed over. Barange and Hampton (1994) observed that the in situ target strength of
Cape horse mackerel (Trachurus trachurus capensis, Carangidae) shifted downwards by
about 12 dB during trawling, indicating increased tilt angle and downward avoidance.
There is increasing evidence that the sampling principle is an oversimplification

(Engas, 1994; God, 1994). Main and Sangster (1981) observed that cod search towards
bottom when approached by a bottom trawl, haddock rise and may escape over the
headline in substantial numbers, while whiting (Merlangius merlangus, Gadidae) tend to
aggregate more in the middle of the trawl opening. Such species-dependent reactions
may clearly result in a species composition in the trawl catch that is not representative
for the area sampled. The reactions of cod differ among size groups so that the smallest
fish escape under the group gear of ordinary sampling trawls and are grossly

underestimated in catches (Engas and God, 1989; Walsh, 1991). Buerkle and
Stephenson (1990) obtained substantial differences in the length distributions of herring
depending on whether the sampling was conducted by a pelagic trawl or a bottom
trawl. As the herring occurred in a dense aggregation extending up from the bottom,
this result was probably caused partly by different vertical distributions but probably
also by a fish-size-dependent reaction to the gears. Gear avoidance seems mainly to be
elicited by visual stimuli, and the reactions decrease at night-time (Glass and Wardle,
1989). In case of mixed species communities that are spatially structured, the number
and location of identification hauls may significantly affect the acoustic biomass
estimates of the respective species (Masse and Retiere, 1995).

`
During surveying in areas with high plankton productivity, the recordings may be so
dominated by contributions from small organisms that allocation of integrated echo
intensity to fish is difficult. However, it is usually possible to sort out recordings of fish
schools, and the difficulty may be partially overcome by thresholding (MacLennan and
Simmonds, 1992).
Acoustic classification of fish species and size
If fish species and size could be determined with reasonable confidence from
information inherent in the reflected acoustic signal, the time-consuming, costly and
often troublesome identification by trawling would not be necessary. Efforts to identify
species and fish size by underwater acoustics indicate that such discrimination is
possible, but difficult. Rose and Leggett (1988) found that the probability distribution of
echo amplitude from fish schools insonified by narrowband echosounder could be used
to discriminate species such as cod, herring and mackerel. On the other hand, Scalabrin
and Lurton (1994) concluded that narrowband, echo amplitude probability distribution
functions of schools of anchovy, blue whiting (Micromesistius poutassou, Gadidae), horse

18

Misund

mackerel (Trachurus trachurus, Carangidae) and sardine contained little information


useful for species discrimination. Extending the analysis to echogram features,
probability density function of echo amplitudes and spectral characteristics of the echo
envelopes of schools recorded by a narrowband transducer, Scalabrin et al. (1996)
obtained an overall correct species classification of only 57%. The low discrimination
was claimed to be a consequence of the low resolution of the narrowband echosounder,
and due to regional and seasonal variation in the shoaling behaviour of the species
investigated. When limiting the analysis to a short season and a specific region, the
species were discriminated with up to 98% confidence (Scalabrin et al., 1996). Similar,
precise species classification has been presented by use of neural network analysis of
echogram features of schools of anchovy, horse mackerel and sardine recorded with a
narrowband echosounder in a limited area in the Mediterranean (Haralabous and
Georgakarakos, 1996). Using eight frequencies in the range 27 to 54 kHz, Simmonds and
Armstrong (1990) identified herring, mackerel and the gadoid species cod and saithe
with 90% confidence. The separation between the gadoid species was more difficult and
they were identified to the correct species with about 7080% confidence. In a similar
experiment, Simmonds et al. (1996) obtained up to 95% confidence in identification of
the gadoids cod, haddock, and saithe and other species like mackerel and horse mackerel
when analysing the frequency spectra by classical discriminant analysis or by neural
network. The confidence of the identification decreased the lower the sample size.
Similarly, Zakharia et al. (1996) obtained up to 75% correct classification of anchovy,
sardine and horse mackerel schools in the Bay of Biscay using a wideband echo sounder
and neural network classification. Holliday (1977) demonstrated that fish can be
classified to size by use of the resonance frequency of the swim bladder. Barange et al.
(1994) concluded that target strength information can be useful in studies of small-scale
community structure and dynamics of fish populations in low-density, multispecific
assemblages. Richards et al. (1991) classified rockfish assemblages according to bottom
habitat category by use of echogram features such as mean volume density, dispersion
and distance off bottom.
Spatial variability
The whole distribution area of a fish stock must be surveyed if a reliable acoustic
abundance estimate is to be obtained. Methods to design and conduct acoustic surveys
are reviewed and discussed by Simmonds et al. (1992). Usually, fish stocks are surveyed
either by exploratory zigzag transects or systematically by equidistant, parallel transects.
If the distribution of the stock is reasonably well known a priori, the survey can be
designed with parallel, random transects so that both the mean abundance estimate and
the corresponding variance can be calculated using classical statistics (Jolly and
Hampton, 1990). However, a more precise estimate of the biomass in an area will be
obtained by a stratified random survey or a systematic survey (Simmonds and Fryer,
1996). Fish stocks must preferably be surveyed during a limited time period to obtain
abundance estimates that are as synoptic as possible. In practical work this intention can
be difficult to follow when surveying widely distributed fish stocks, such as the North
Sea herring which may take 34 weeks to map even by coordinated surveys with several
participants (Bailey and Simmonds, 1990).
Acoustic abundance estimates may be significantly biased if the fish are migrating
during surveys (MacLennan and Simmonds, 1992). The distribution area will be over-

Underwater acoustics in marine fisheries

19

or underestimated if the fish are migrating along or against the direction in which the
survey is progressing, respectively. However, it is possible to correct abundance
estimates for the migration effect if the migration speed of the fish along or against the
progression speed of the survey is known (MacLennan and Simmonds, 1992). For
schooling fish, Hafsteinsson and Misund (1995) have developed a method to record the
migration speed during acoustic surveys by use of sonar.
Density distribution. Some fish species live mostly individually, others aggregate in
shoals, while the highest densities are formed by species that organize themselves into
schools. The aggregative behaviour induces a large dispersion and skewness in the
distribution function of the area backscattering strength (sA ) recorded during acoustic
surveys. This is especially evident when the fish assessed are schooling (Aglen, 1983).
The confidence limits of the estimates when using classical statistics may therefore be
very large, but according to the central limit theorem the two first moments (mean and
variance) of the distribution function are still finite.
However, there exist examples where the distribution function of the data sampled do
not have these properties, but instead an approximately infinite variance. In such cases
an increase of the sampling rate has a limited effect on the precision of the estimator.
The distribution function of data from 18 acoustic surveys in tropical areas resembles
that of a Pareto distribution with an infinite variance (Freon et al., 1993). Even though

curve-fitting using the likelihood approach was not perfect for the highest values of
biomass, these results suggest that the arithmetic mean of the samples does not give a
good estimate of the population. However, it is the experience of many surveys
conducted to give absolute population estimates that there is a fairly good precision. By
use of the delta distribution, MacLennan and MacKenzie (1988) estimated a precision
of about 25% in acoustic estimates of North Sea herring.
Aglen (1983) showed that the variability in acoustic survey estimates decreases with
increasing degree of coverage, and proposed that the coefficient of variation of an
acoustic abundance estimate is proportional to the inverse square root of the degree of
coverage divided by the square root of the area covered. Gerlotto and Stequert (1983)

also demonstrated lower variability with shorter distance among transects. Hansson
(1993) found that the average coefficient of variation between replicates from one night
along a single transect when recording herring in a relatively closed Baltic coastal area
was 20%. For replicates of a transect from different nights, the average coefficient of
variation increased to 29% due to migration.
Calculating the variance of acoustic abundance estimates obtained by conventional
surveying is usually not straightforward because serial correlation of the recorded
backscattering strengths (sA ) due to patchily distributed fish violates the use of classical
statistics (Jolly and Hampton, 1990). It is possible that estimation of variance from
systematic surveys may be done by use of geostatistics, which incorporates the
observed spatial structure into the calculation of variance (Petitgas, 1993; Simard et al.,
1993; Williamson and Traynor, 1996). There is already software available to calculate
the variance of acoustic survey estimates by use of geostatistics (Petitgas and Prampart,
1993).

Replicability of survey estimates. The very skewed distributions of acoustic density


measurements of pelagic stocks indicate that there may be substantial variation in

20

Misund

population estimates between surveys. In accordance with this, Buerkle and Stephenson
(1990) found large variations in biomass estimates from repeated night-time surveys of
herring in Chedabucto Bay, Canada. They claimed that the variations were caused by a
great dynamic in the aggregations of the population. To some extent this variation may
also have been caused by the application of fixed transects so that parts of the stock may
have been out of the area from one survey to another. Nevertheless, Strmme and
Stersdal (1987) observed a variation of less than 20% when repeating surveys of a
length from about 24 hours to 7 days on pelagic stocks off Senegal and Morocco.
Vilhjalmsson et al. (1983) showed a high replicability of acoustic estimates of Icelandic

capelin, and noticeable deviations were reasonably explained by ice cover or surface
schooling. Repeated acoustic surveys of Norwegian spring-spawning herring hibernating
in fjords in northern Norway show that the abundance estimates may vary by factors of
1.062.14 for surveys 12 months apart (Table 1).
Comparison with other independent methods
The echo integration technique has become a convenient and widely applied method for
abundance estimation of commercially important fish stocks. However, as outlined in this
review, the method is still afflicted with uncertainties regarding the in situ backscattering
cross section of the fish, dead zones, behavioural effects, and considerable uncertainty
connected to the spatial extrapolation of the recordings. To be validated, acoustic
abundance estimates should therefore be compared with other independent estimates of
the same fish stocks. Comparison with assessment by virtual population analysis (VPA)
has little confidence owing to covariation if the former have been used to tune the latter.
Nevertheless, abundance estimation of marine fish stocks is expensive, and there exist
few examples of independent comparison with acoustic and other independent estimates
of fish stocks. Bailey and Simmonds (1990) found that acoustic surveys of North Sea
herring provided similar indications of changes in abundance to those derived from
herring larvae surveys or VPA based on catches. Hampton (1996) showed strong
correlations between acoustic and egg-production estimates of the spawning population
of anchovy (Engraulis capensis [ japonicus], Engraulidae) off South Africa in the
period 19851994. The mean of the acoustic spawner biomass estimates agreed to within
10% of the egg-production mean when using the target strength equation of Barange et
al. (1996). When comparing acoustic and bottom-trawl survey indices for cod and
haddock in the Barents Sea, God (1994) noted low correlation and considerable amongyear variation for the two indices. The inconsistencies were explained by varying
availability of these semipelagic species to being captured by trawl in the bottom zone or
to being recorded by echo integration in the pelagic zone. Carscadden et al. (1994) found
a certain correspondence between acoustic abundance estimates, aerial survey indices and

Table 1. Acoustic abundance estimates (N) corrected for acoustic shadowing and with geostatistical
variance (N/N) of Norwegian spring-spawning herring wintering in fjords in northern Norway (from
Rttingen et al., 1994; Foote and Rttingen, 1995; Foote et al., 1996)
Dec. 1993
N (109 )
N/N

Jan. 1994

Dec. 1994

Jan. 1995

Nov. 1995

Jan. 1996

13.2

14.6

15.2
0.08

16.2
0.07

15.0
0.05

32.1
0.06

Underwater acoustics in marine fisheries

21

catch rates of capelin at the coast of Newfoundland, and expressed most confidence in
the acoustic estimates. Hansson and Rudstam (1995) found a weak, but significant
correlation between gillnet catch rates and acoustic abundance estimates of herring in
archipelago areas of the northern Baltic Sea. These authors argue that acoustics may give
reliable estimates of absolute biomass, and that gillnet catch rates may be a biased
indicator of changes in fish abundance due to visibility and temperature effects.
THE ECHO COUNTING METHOD

If the fish are distributed individually and in a sufficiently low density it is possible to
estimate their density by counting the number of targets and dividing by the sampling
volume. The total biomass of the area investigated is then found by multiplying the
recorded fish density by the volume of water in the area (MacLennan and Simmonds,
1992).
A prerequisite for the echo counting method is that individual fish of the same size
have the same target strength independent of range (R). This is obtained by applying a
40 log R - time-varied gain (TVG) function which compensates for the transmission loss
due to geometrical spread. The sonar equation (1) then takes the form:
2
2 R

.R

.(

.e

4 ). N

b2 (, ) d

I 0 . (c 2) .

(5)

For successful application of the method it is necessary to have proper instrumentation


and methodology to identify and count returning echoes from single fish. The received
signal must pass a threshold criterion to eliminate noise, and then an identification filter
that accepts echoes originating from single fish and rejects aggregated signals from
several fish that occur within the pulse volume. This is done by setting criteria for the
duration, amplitude and curvature of the received signals. For fish to occur as separate
echoes, they must be more than half a pulselength apart in the direction of the acoustic
beam, and more than the width of the main lobe in the direction perpendicular to the
beam. The sampling volume of the emitted beam depends on the target strength of the
fish, and it is therefore necessary to have continuous recording of the target strength.
This can be done indirectly for a single-beam sounder by removing the beam pattern
effect according to the methods of Craig and Forbes (1969), Ehrenberg (1972) or Clay
(1982), or directly for a split-beam or dual-beam sounder.
T H E S O NA R M E T H O D

Horizontal guided sonar is an efficient tool for recording fish schooling close to the
surface, and may be used scientifically to count fish schools and estimate fish abundance
(Smith, 1970; Lamboeuf et al., 1983). With the intention of using horizontal guided
sonar for acoustic abundance estimation a computerized system for automatic detection,
counting and sizing of fish schools was developed in the 1970s (Hewitt et al., 1976).
However, the method has seldom been used because it has been difficult to establish
factors for converting the parameters measured by the sonar (number of schools, school
area or school target strength) to fish biomass. Now the method is under development,
based on the use of high-resolution sonar (Gerlotto et al., 1994; Misund et al., 1995, Fig.
2), computer-based school detection (Misund et al., 1994), and consistent relationships
between the geometric dimensions (area or volume) and biomass of fish schools (Misund

22

Misund

et al., 1992) which enable conversion of sonar measurements of school dimensions to


school biomass.
A fascinating application of underwater acoustics to record fish shoals is Weston and
Andrews (1990) use of low-frequency (about 12 kHz), long-range (up to 37 km),
fixed sonar to observe sardine shoals in the Bristol Channel. This method should be
further developed for detecting the presence of fish shoals on fishing grounds and for
studies of fish shoal migrations from sites onshore or on offshore platforms.
Fish behaviour and distribution studies
Echosounders and sonars have been vital for mapping the geographical distribution of
many economically important fish stocks, their migration routes and their relationship to
oceanographic features (MacLennan and Simmonds, 1992). The distribution and
migrations horizontally and vertically of larger fish aggregations or populations can be
mapped by echosounding, and the movements of fish schools can be quantified by use of
modern sonars (Misund, 1994). Movements of individual fish can be tracked by acoustic
tags (Arnold et al., 1990), by split-beam echosounders (Ona, 1994), or by echo trace
analysis on single-beam echosounders (Furusawa and Miyanohana, 1988).
M A P P I N G O F F I S H M OV E M E N T S B Y AC O U S T I C TAG S

Fish behaviour in the wild has been studied for more than 30 years by use of small
acoustic tags attached to the body of individual fish (Urquhart and Stewart, 1993). Such
tags are claimed not to affect the behaviour of the observed fish (Arnold and Holford,
1978). Unidirectional, constantly transmitting tags have been used to map the coastal
migration of salmon from a following vessel with a directional receiver (Smith et al.,
1981). At the Fisheries Laboratory, Lowestoft, England, a system based on small,
transponding tags (Mitson and Storeton West, 1971) that could be detected by a highresolution sector-scanning sonar (Voglis and Cook, 1966; Holley et al., 1975) has been
developed to study several aspects of fish behaviour (Arnold et al., 1990). When using
this system to observe the behaviour of plaice (Pleuronectes platessa, Pleuronectidae) in
relation to a Granton otter trawl, Harden Jones et al. (1977) found that the fish tend to be
herded by the bridles so that 70% of the plaice that occurred between the trawl doors
ended up in the path of the trawl. By using the system on individual cod, spotted dogfish
(Scyliorhinus canicula, Scyliorhinidae) plaice, silver eels (Anguilla anguilla, Anguillidae) and sole (Solea solea, Soleidae), several different mechanisms of selective tidalstream transport for feeding, spawning and seasonal migrations have been discovered
(Greer Walker et al., 1978; Arnold and Metcalfe, 1989; Arnold et al., 1994). Such
studies have also shown that cod can perform rapid vertical migrations beyond the range
of its neutral buoyancy, which is controlled by secretion/resorption of gas in the
swimbladder (Arnold et al., 1994). Such migrations probably affect the backscattering
cross section of the fish through changes both in tilt angle and in swimbladder volume
(Arnold et al., 1994).
R E C O R D I N G O F F I S H B E H AV I O U R A N D D I S T R I B U T I O N B Y E C H O S O U N D E R S A N D
S O NA R

The migration routes of the Atlanto-Scandian herring in the north-eastern Atlantic were
mapped by use of echosounder and sonar in the early 1950s (Devold, 1950). By echo

Underwater acoustics in marine fisheries

23

surveying and trawling, Rose (1993) observed that in late winter, cod off Newfoundland
migrated towards the coast in the warmer deep channels between the banks and spawned
on the way. Rose et al. (1995) related the migration speeds of acoustically detected cod
to the currents on the north-east Newfoundland Shelf. Walsh et al. (1995) used
echosounding to map the migrations of mackerel off Scotland, as did Castonguay and
Gilbert (1995) to study the effects of tidal streams on the migration of mackerel off
eastern Canada. Combining echo recordings, satellite images, aerial observations and
commercial catch data, Tameishi et al. (1996) concluded that warm, plankton-rich
streamers play an important role in the northward migration of Japanese sardine
(Sardinops melanostictus, Clupeidae).
Using echo integration recordings to investigate the relationship between walleye
pollock, water depth and temperature in the Bering Sea, Swartzman et al. (1995) found
the highest abundance in areas having a 70130 m depth range and where the
temperature at 50 m depth was close to 2.5 8C. According to Swartzman et al. (1994),
pollock remained below the thermocline when midwater temperatures were below 0 8C,
and fish above the thermocline were largely age-0 pollock while adults predominated
below the thermocline. Castillo et al. (1996) observed high annual and seasonal
variability in the distribution of anchovy (Engraulis japonicus, Engraulidae), jack
mackerel (Trachurus murphyi, Carangidae) and sardine (Sardinops sagax, Clupeidae) off
northern Chile, and anchovy seemed to remain in salinity and temperature ranges that
were lower than observed for the other species. Samb and Demarq (1989) found that
fish aggregated in the periphery of the upwelling area off the south coast of Senegal,
but in the central upwelling area off the north coast of Senegal. The observed
differences were probably related to differences in upwelling intensity and species
composition. Boudreau (1992) observed that most of the spawning biomass of haddock
was , 2.5 m off bottom, and that the fish density was highest during daytime, and
declined the larger the size of the fish. Echosounding has also been used to show that
fish tend to concentrate around oil platforms offshore (Gerlotto et al., 1989; Stanley
and Wilson, 1996).
Considerable information about the formation and characteristics of fish schools in
the wild has been obtained by echosounder and sonar recordings. Acoustic observations
at sea indicate that pelagic fish generally aggregate in distinct schools at dawn and
disperse into more-loosely organized shoals at dusk (Blaxter and Hunter, 1982). By use
of a computer-based, post-processing system with a special school integration software
package, Freon et al. (1996) found that small, but loose and irregularly shaped schools

could be recorded at night in the Mediterranean. During daytime the schools had a high
packing density, tended to be larger in area, and were rather regularly shaped. Misund
(1993) found that the packing density of herring, sprat and saithe shools was inversely
proportional to the cube of the fish size. When improving the resolution of conventional
echo integration, Freon et al. (1992) and Misund (1993) found that the packing density

within clupeid schools varied by a factor of about 100, and that both high-density
regions and empty lacunas could be detected within the schools. Similar
heterogeneous packing within fish shoals was observed by Cushing (1977) by use of
a high-resolution sonar.
Hara (1985) observed that the shape of Japanese sardine was size dependent so that
small schools tend to be oval while elongate or crescent-like schools were 10 m in
width and 100200 m in length. Misund (1993) found that herring, sprat and saithe

24

Misund

schools tend to be spherical in mid-water and discoidal when close to surface or sea
bottom. Similarly, Ohshimo (1996) observed that the horizontal extent was about five
times the vertical extent of anchovy, Engraulis japonicus, schools that were distributed
in shallow water (,50 m depth) in the China Sea. Marchal and Petitgas (1993) found
that the biomass distribution of echo-integrated schools in the Gulf of Cariaco,
Venezuela, was highly skewed because there were numerous small schools and only a
few larger ones. Scalabrin and Masse (1993) found an annual variation in shoal size

and species-dependent differences in shoal characteristics between bathymetric zones in


the Bay of Biscay. In this area horse mackerel were found close to bottom, while
anchovy occurred about 16 m off bottom (Masse et al., 1996). When sprat mixed with

anchovy, the shape of the schools was modified (Masse et al., 1996). The size of the

herring, sprat and saithe schools in the North Sea and along the Norwegian coast varied
from a few hundred to several million fish (Misund, 1993). Regional and seasonal
differences in packing density, size and geometry among fish schools can probably be
related to variations in motivational state of the fish, predatory regime and other
environmental features. Petitgas and Levenez (1996) found that schools off Senegal

occurred in clusters, and that the dimensions and number of clusters relative to the area
where they were found stayed constant with varying total biomass. However, the area in
which fish were present and the occurrence of dense schools varied according to the
total biomass. Sonar observations (Fig. 2) of herring schools on feeding migration in
the Norwegian Sea showed a dynamic regime with frequent approaching, joining,
leaving and splitting of schools (Pitcher et al., 1996). Adaptive responses to attacks by
predators caused frequent modifications to the structure of the schools so that most
herring probably were able to continue feeding while retaining the risk-dilution
advantages of schooling.

Summary and prospects


Using underwater acoustics to measure fish density at sea is attempting to achieve
accurate and precise values through very variable signals. The variations are caused by
the instrument characteristics and settings, the beam geometry, and the reflection
properties of the fish especially because of its orientation, but also because of its
swimming depth, maturation stage, stomach fullness and fat content, and because of the
swimming behaviour of the fish when being recorded: whether it is schooling, shoaling
or swimming individually, and if it is avoiding the approaching survey vessel. In
addition, acoustic fish density estimates will be affected by how well the distribution area
of the stock is covered by the survey, and the spatial variability of the fish within the
survey area. The acoustic recordings have to be identified by fishing, usually by trawling,
and the reaction to the sampling gear of different species may therefore also affect the
acoustic fish density estimate.
Nevertheless, despite all these sources of error and variability, underwater acoustics
can be used to make absolute measurements of fish abundance (Foote and Knudsen,
1994), and the methodology is now used for providing fishery-independent estimates of
many economically important fish stocks around the world. Still, there is great potential
for further development of underwater acoustics to provide more accurate and precise
estimates, with reliable variance of fish abundance. More accurate and sophisticated

Underwater acoustics in marine fisheries

25

instruments must be developed that can extract more information about fish abundance,
species and size from the reflected signals. Multibeam, multifrequency sounders that
can extract the information within the different beams with high resolution and make
comparisons of the information between the beams for different frequencies may be a
solution when searching for instruments that can measure absolute fish density as well
as identify fish species and size. The reflecting properties of the fish and how these
properties change due to behaviour and physiology are still far from satisfactorily
explored. In future surveys, dynamic backscattering cross section functions that account
for physiological and behavioural influence will probably have to be applied when
converting area backscattering strength (sA ) to fish density. How the accuracy, precision
and variance of acoustic survey estimates will be influenced by spatial variability of the
fish stock surveyed will need special attention in future research.
Many economically important semipelagic fishes like cod and pollock are surveyed
by both bottom trawl and acoustics, providing two separate indices of abundance.
Pelagic schooling species that sustain large fisheries occasionally occur close to the
surface, and partially or completely in the upper dead zone of the echosounder. To
improve the assessments, point estimates of fish density must be provided by combining
the bottom trawl and acoustic recordings of semipelagic fish, and similarly by
combining sonar and echo integration estimates of near-surface, schooling fish (Fig. 3).
Principally this can be done by observing the behaviour of the fish during recording so
that the availability to the recording method (sonar, echosounder or bottom trawl) can
be determined and used to obtain combined estimates of fish density.
In many areas with high species diversity, identification of acoustic recordings is still
complicated and uncertain. Further research must therefore be devoted to studying the
reactions of fish to sampling gears and to improving the efficiency of the gears.
Similarly, the possibility of identifying the fish on the basis of species-specific
information inherent in the reflected signals must also be explored further.
The behaviour of fish in the wild is still rather unexplored, and further development
of underwater acoustics is necessary to gain more knowledge of how fish behave in the
sea. This must be done through further developments of acoustic tags for studying
various aspects of individual fish behaviour, data-storage tags for gathering information
about movements of individual fish over longer time periods (Arnold and Holford,
1995), and sophisticated, high-resolution sonars to study the behaviour of schooling
fish.
Underwater acoustics has been applied regularly in fisheries and fisheries research
for less than 50 years. Undoubtedly there is still commercial potential for developing
improved and sophisticated acoustic instruments for fisheries application. Such
instruments must not only help fishers in locating the fish, but also give them
information about fish species, fish size and abundance prior to catching. In addition to
improving the efficiency, such instruments will help fishers to adopt a more responsible
attitude when fishing. For example, if fishers are given a reliable estimate of school
size, they may avoid purse seining on schools that are too big and can cause net rupture
and substantial unaccounted mortality. If given information on fish species and fish
size, they can avoid fishing in areas with a lot of non-target species or undersized fish.
Then underwater acoustics can contribute to developing sustainable fisheries, both
through the methods for abundance estimation and behaviour studies, and by enabling
more responsible fishing.

26

Misund

Acknowledgements
I am grateful to two distinguished referees for constructive and helpful comments that
enabled a substantial improvement of this review.
References
Aglen, A. (1983) Random errors of acoustic fish abundance estimates in relation to the survey grid
density applied. FAO Fish. Rep. 300, 2938.
Aglen, A. (1994) Sources of error in acoustic estimation of fish abundance. In Ferno, A. and Olsen,

S., eds. Marine Fish Behaviour in Capture and Abundance Estimation. Oxford: Fishing News
Books, pp. 107133.
Anon. (1992) Dolphins and the Tuna Industry. Washington, DC: National Academy Press. 176 pp.
Appenzeller, A.R. and Leggett, W.C. (1992) Bias in hydroacoustic estimates of fish abundance due to
acoustic shadowing: evidence from daynight surveys of vertically migrating fish. Can. J. Fish.
Aquat. Sci., 49, 217989.
Arnold, G.P. and Holford, B.H. (1978) The physical effects of an acoustic tag on the swimming
performance of plaice and cod. J. Cons. Int. Explor. Mer 38, 189200.
Arnold, G.P. and Holford, B.H. (1995) A computer simulation model for predicting rates and scales of
movement of demersal fish on the European continental shelf. ICES J. Mar. Sci. 52, 98190.
Arnold, G.P. and Metcalfe, J.D. (1989) Acoustic telemetry: progress and potential in understanding fish
behaviour. Proc. Inst. Acoust. 11, 96103.
Arnold, G.P., Greer Walker, M. and Holford, B.H. (1990) Fish behaviour: achievements and potential

of high-resolution sector-scanning sonar. Rapp. P.-v. Reun. Cons. Int. Explor. Mer 189, 11222.
Arnold, G.P., Greer Walker, M., Emerson, L.S. and Holford, B.H. (1994) Movements of cod (Gadus
morhua L.) in relation to the tidal streams in the southern North Sea. ICES J. Mar. Sci. 51,
20732.
Au, W.W.L. (1993) The Sonar of Dolphins. New York: Springer-Verlag. 277 pp.
Bailey, R.S. and Simmonds, E.J. (1990) The use of acoustic surveys in the assessment of the North

Sea herring stock and a comparison with other methods. Rapp P.-v. Reun. Cons. Int. Explor. Mer
189, 917.
Balchen, J.G. (1984) Recent progress in the control of fish behaviour. Modelling, Ident. Control 52,
11321.
Balls, R. (1948) Herring fishing with the echometer. J. Cons. Perm. Int. Explor. Mer 15, 193206.
Barange, M. and Hampton, I. (1994) Influence of trawling on in situ estimates of Cape horse mackerel
(Trachurus trachurus capensis) target strength. ICES J. Mar. Sci. 51, 1216.
Barange, M., Hampton, I. and Soule, M. (1996) Empirical determination of in situ target strengths of
three loosely aggregated pelagic fish species. ICES J. Mar. Sci. 53, 22532.
Barange, M., Hampton, I., Pillar, S.C. and Soule, M.A. (1994) Determination of composition and
vertical structure of fish communities using in situ measurements of acoustic target strength. Can.
J. Fish. Aquat. Sci. 51, 99109.
Beltestad, A.K. (1974) Feeding behaviour, vertical migration and schooling among O-group herring
(Clupea harengus L.) in relation to light intensity. Thesis, Univ. Bergen. 80 pp. (in Norwegian,
unpublished).

Blaxter, J.H.S. and Batty, R.S. (1990) Swimbladder behaviour and target strength. Rapp. P.-v. Reun.
Cons. Int. Explor. Mer 189, 23344.
Blaxter, J.H.S. and Hunter, J. (1982) The biology of clupeoid fishes. Adv. Mar. Biol. 20, 1223.
Blaxter, J.H.S. and Tytler, P. (1978) Physiology and function of the swimbladder. Adv. Comp. Physiol.
Biochem. 7, 31167.
Bodholdt, H. (1982) A multi-beam sonar for fish school observations. ICES/FAO Symposium on
Fisheries Acoustics, Bergen, Norway, 2124 June. Doc. no. 55. (unpublished).

Underwater acoustics in marine fisheries

27

Bodholdt, H. and Olsen, K. (1977) Computer-generated display of an underwater situation:

applications in fish behaviour studies. Rapp. P.-v. Reun. Cons. Int. Explor. Mer 170, 315.
Bodholt, H., Ness, H. and Solli, H. (1989) A new echo-sounder system. Proc. Inst. Acoust. 11,
12330.
Boudreau, P. (1992) Acoustic observations of patterns of aggregation in haddock (Melanogrammus
aeglefinus) and their significance to production and catch. Can. J. Fish. Aquat. Sci. 49, 2331.
Brawn, V
.M. (1962) Physical properties and hydrostatic function of the swimbladder of herring
(Clupea harengus L.). J. Fish. Res. Bd Can. 19, 63556.
Buerkle, U. (1983) First look at herring distribution with a bottom referencing underwater towed
instrumentation vehicle BRUTIV. FAO Fish. Rep. 300, 12530.
Buerkle, U. and Stephenson, R.L. (1990) Herring school dynamics and its impact on acoustic
abundance estimation. Proceedings of the International Herring Symposium, Anchorage, Alaska,
USA, October 2325, 1990, pp. 185208.
Burns, D.R., Qeen, C.B., Sisk, H., Mullarkey, W. and Chivers, R.C. (1989) Rapid and convenient
acoustic sea-bed discrimination for fisheries applications. Proc. Inst. Acoust. 11, 16978.
Carscadden, J., Nakashima, B. and Miller, D.S. (1994) An evaluation of trends in abundance of
capelin (Mallotus villosus) from acoustics, aerial surveys and catch rates in NAFO division 3L,
198289. J. Northw. Atl. Fish. Sci. 17, 4557.
Castillo, J., Barbieri, M.A. and Gonzalez, A. (1996) Relationship between sea surface temperature,
salinity, and pelagic fish distribution off Chile. ICES. J. Mar. Sci. 53, 13946.
Castonguay, M. and Gilbert, D. (1985) Effects of tidal streams on migrating Atlantic mackerel,
Scomber scombrus L. ICES J. Mar. Sci. 52, 94154.
Clay, C.S. (1982) Deconvolution of the fish scattering PDF from the echo PDF for a single transducer
sonar. J. Acoust. Soc. Am. 73, 198994.
Clay, C.S. and Medwin, H. (1977) Acoustical Oceanography: Principles and Application. New York:
Wiley. 544 pp.
Craig, R.E. and Forbes, S.T. (1969) A sonar for fish counting. FiskDir. Skr. Ser. Havunders. 15,
21019.

Cushing, D. (1977) Observations of fish schools with the ARL scanner. Rapp. P.-v. Reun. Cons. Int.
Explor. Mer 170, 1520.
Dalen, J. and Lvik, A. (1981) The influence of wind-induced bubbles on echo-integration surveys.
J. Acoust. Soc. Am. 69, 16539.
Dalen, J. and Nakken, O.M. (1983) On the application of the echo integration method. ICES CM
1983/B:19, 30 pp.
Dawson, J.J. and Brooks, T.J. (1989) An innovative acoustic signal processor for fisheries science.
Proc. Inst. Acoust. 11, 13140.
Degnbol, P., Lassen, H. and Staer, K.J. (1985) In-situ determination of target strength of herring and
sprat at 38 and 120 kHz. Dana 5, 4554.
Devold, F. (1950) Herring cruise with G.O. Sars in the Norwegian Sea July 5th August 24th
1950. Fiskets Gang 36, 4646. (in Norwegian).
Diner, N. and Masse, J. (1987) Fish school behaviour during echo survey observed by acoustic device.
International Symposium on Fisheries Acoustics, 2124 June, Seattle, USA.
Dragesund, O. and Olsen, S. (1965) On the possibility of estimating year-class strength by measuring
echo-abundance of O-group fish. FiskDir. Skr. Ser. Havunders. 13, 4775.
Edwards, J.I. and Armstrong, F. (1983) Measurements of the target strength of live herring and
mackerel. FAO Fish. Rep. 300, 6977.
Edwards, J.I., Armstrong, F., Magurran, A.E. and Pitcher, T.J. (1984) Herring, mackerel and sprat
target strength experiments with behavioural observations. ICES CM 1984/B:34, 21 pp.
Ehrenberg, J.E. (1972) A method for extracting the fish target strength distribution from acoustic
echoes. Conference 1315 September 1972, Newport, Rhode Island, USA. Proc. 1972 IEEE
Conf. Eng. Ocean Environ. New York: IEEE, pp. 614.

28

Misund

Ehrenberg, J.E. (1983) A review of in situ target strength estimation techniques. FAO Fish. Rep. 300,
8590.

Engas, A.E. (1994) The effects of trawl performance and fish behaviour on the catching efficiency of
demersal sampling trawls. In Ferno, A. and Olsen, S., eds. Marine Fish Behaviour in Capture

and Abundance Estimation. Oxford: Fishing News Books, pp. 4568.

Engas, A.E. and God, O.R. (1989) Escape of fish under the fishing line of a Norwegian sampling
trawl and its influence on survey results. J. Cons. Int. Explor. Mer 45, 26976.

Engas, A.E., Misund, O.A., Soldal, A.V Horvei, B. and Solstad, A. (1995) Reactions of penned
.,
herring to playback of original, frequency filtered and time-smoothed vessel sound. Fish. Res. 22,
24354.
Foote, K.G. (1978) Analysis of empirical observations on the scattering of sound by encaged
aggregations of fish. FiskDir. Skr. Ser. Havunders. 16, 42356.
Foote, K.G. (1980a) Effects of fish behaviour on echo energy: the need for measurements of
orientation distributions. J. Cons. Int. Explor. Mer 39, 193201.
Foote, K.G. (1980b) Importance of the swimbladder in acoustic scattering by fish: a comparison of
gadoid and mackerel target strengths. J. Acoust. Soc. Am. 67, 20849.
Foote, K.G. (1981) Evidence for the influence of fish behaviour on echo energy. In Soumala, J.B., ed.
Meeting on Hydroacoustical Methods for the Estimation of Marine Fish Populations, 2529 June
1979, Vol. 2. Cambridge, MA: Charles Stark Draper Laboratory Inc., pp. 20128.
Foote, K.G. (1983) Linearity of fisheries acoustics, with addition theorems. J. Acoust. Soc. Am. 73,
193240.
Foote, K.G. (1985) Rather-high-frequency sound scattering by swimbladdered fish. J. Acoust. Soc. Am.
78, 688700.
Foote, K.G. (1987) Fish target strengths for use in echo integrator surveys. J. Acoust. Soc. Am. 82,
9817.
Foote, K.G. (1990) Correcting acoustic measurements of scatterer density for extinction. J. Acoust.
Soc. Am. 88, 15436.
Foote, K.G. and Knudsen, H.P. (1994) Physical measurements with modern echo integrators. J. Acoust.
Soc. Jpn 15, 3935.
Foote, K.G. and Nakken, O. (1978) Dorsal-aspect target strength functions of six fishes at two
ultrasonic frequencies. Fisken og Havet. Ser. B, 1978, Institute of Marine Research, Bergen, 95
pp.
Foote, K.G. and Rttingen, I. (1995) Acoustic assessment of Norwegian spring spawning herring in
the wintering area, December 1994 and January 1995. ICES C.M. 1995/H:9, 11 pp.
Foote, K.G., Knudsen, H.P., Vestnes, G., MacLennan, D.N. and Simmonds, E.J. (1987) Calibration of
acoustic instruments for fish density estimation: a practical guide. Int. Coun. Explor. Sea Coop.
Res. Rep. No. 144. 57 pp.
Foote, K.G., Ona, E. and Toresen, R. (1992) Determining the extinction cross section of aggregating
fish. J. Acoust. Soc. Am. 91, 19839.

Foote, K.G., Ostrowski, M., Rttingen, I., Engas, A., Hansen, K., Hiis Hauge, K., Skeide, R., Slotte,
A. and Torgersen, . (1996) Acoustic abundance estimation of the stock of Norwegian spring
spawning herring, winter 19951996. ICES C.M. 1996/H:33, 13 pp.
Forbes, S.T. and Nakken, O. (1972) Manual of Methods for Fisheries Resource Survey and Appraisal.
Part 2. The Use of Acoustic Instruments for Fish Detection and Abundance Estimation (FAO
Man. Fish. Sci. 5) Rome: FAO. 138 pp.
Freon, P., Gerlotto, F. and Soria, M. (1990) Evaluation of the influence of vessel noise on fish

distribution as observed using alternatively motor and sails aboard a survey vessel. ICES FAST
Working Group, Rostock, 1990.
Freon, P., Gerlotto, F. and Soria, M. (1992) Changes in school structure according to external stimuli:

description and influence on acoustic assessment. Fish. Res. 15, 4566.


Freon, P., Soria, M., Mullon, C. and Gerlotto, F. (1993) Diurnal variation in fish density estimate

Underwater acoustics in marine fisheries

29

during acoustic surveys in relation to spatial distribution and avoidance reaction. Aquat. Liv.
Resour. 6, 22134.
Freon, P., Gerlotto, F. and Soria, M. (1996) Diel variability of school structure with special reference

to transition periods. ICES J. Mar. Sci. 53, 45964.


Furusawa, M. (1988) Prolate spheroidal models for predicting general trends of fish target strength.
J. Acoust. Soc. Jpn 9, 1324.
Furusawa, M. and Miyanohana, Y. (1988) Application of echo-trace analysis to estimation of
behaviour and target strength of fish. J. Acoust. Soc. Jpn 9, 16980.
Gerlotto, F. and Freon, P. (1990) Review of avoidance reactions of tropical fish to a survey vessel.

ICES FAST Working Group. Rostock, 1990 (Unpublished).


Gerlotto, F. and Freon, P. (1992) Some elements of vertical avoidance of fish schools to a vessel

during acoustic surveys. Fish. Res. 14, 2519.


Gerlotto, F. and Stequert, B. (1983) Une methode de simulation pour etudier la distribution des

densites en poissons: application a deux cas reels. FAO Fish. Rep. 300, 27892.

`
Gerlotto, F., Bercy, C. and Bordeau, B. (1989) Echo-integration survey around off-shore oil-extraction
platforms off Cameron: observation of the repulsive effect on fish of some artificially emitted
sounds. Proc. Inst. Acoust. 11, 7988.
Gerlotto, F., Petit, D. and Freon, P. (1990) Influence of the light of a survey vessel on TS distribution.

ICES FAST Working Group, Rostock, 1990 (Unpublished).


Gerlotto, F., Freon, P., Soria, M., Cottais, P.-H. and Ronzier, L. (1994) Exhaustive observation of the

3D schools structure using multibeam side scan sonar: potential use for school classification,
biomass estimation and behaviour studies. ICES CM 1994/B:26, 12pp.
Glass, C. and Wardle, C.S. (1989) Comparisons of the reactions of fish to a trawl gear, at high and
low light intensities. Fish. Res. 7, 24966.
God, O.R. (1994) Factors affecting the reliability of groundfish abundance estimates from bottom
trawl surveys. In Ferno, A. and Olsen, S., eds. Marine Fish Behaviour in Capture and Abundance

Estimation. Oxford: Fishing News Books, pp. 16699.


Goncharov, S.E., Borisenko, E.S. and Pyanov, A. (1989) Jack mackerel schools defence reaction to a
surveying vessel. Proc. Inst. Acoust. 11, 748.
Greer Walker, M., Harden Jones, F.R. and Arnold, G.P. (1978) The movements of plaice (Pleuronectes
platessa L.) tracked in the open sea. J. Cons. Int. Explor. Mer 38, 5886.
Hafsteinsson, M.T. and Misund, O.A. (1995) Recording the migration behaviour of fish schools by
multi-beam sonar during conventional acoustic surveys. ICES J. Mar. Sci. 52, 91524.
Halldorsson, O. and Reynisson, P. (1983) Target strength measurement of herring and capelin in-situ.
ICES CM 1983/H:36, 35 pp.
Hampton, I. (1996) Acoustic and egg-production estimates of South Africa anchovy biomass over a
decade: comparisons, accuracy, and utility. ICES J. Mar. Sci. 53, 493500.
Hansson, S. (1993) Variation in hydroacoustic abundance of pelagic fish. Fish. Res. 16, 20322.
Hansson, S. and Rudstam, L.G. (1995) Gillnet catches as an estimate of fish abundance: a comparison
between vertical gillnet catches and hydroacoustic abundance of Baltic Sea herring (Clupea
harengus) and sprat (Sprattus sprattus). Can. J. Fish. Aquat. Sci. 52, 7583.
Hara, I. (1985) Shape and size of Japanese sardine school in the waters off the southeastern Hokkaido
on the basis of acoustic and aerial surveys. Bull. Jap. Soc. Scient. Fish. 51, 416.
Haralabous, J. and Georgakarakos, S. (1996) Artificial neural networks as a tool for species
identification of fish schools. ICES J. Mar. Sci. 53, 17380.
Harden Jones, F.R. and Scholes, P. (1981) The swimbladder, vertical movements and the target
strength of fish. In Soumala, J.B., ed. Meeting on Hydroacoustical Methods for the Estimation of
Marine Fish Populations, 2529 June 1979, Vol. 2. Cambridge, MA: Charles Stark Draper
Laboratory, pp. 15782.
Harden Jones, F.R., Margetts, A.R., Greer Walker, M. and Scholes, P. (1977) The efficiency of the
Granton otter trawl determined by sector-scanning sonar and acoustic transponding tags. Rapp.

30

Misund

P.-v. Reun. Cons. Int. Explor. Mer 170, 4551.


Hawkins, A.D. (1993) Underwater sound and fish behaviour. In Pitcher, T.J., ed. Behaviour of Teleost
Fishes, 2nd edn. London: Chapman & Hall, pp. 12969.
He, P. and Wardle, C.S. (1986) Tilting behaviour of the Atlantic mackerel, Scomber scombrus, at low
swimming speeds. J. Fish Biol. 29, 22332.
Hewitt, R.P., Smith, P.E. and Brown, J.C. (1976) Development and use of sonar mapping for pelagic
stock assessment in the California Current area. Fish. Bull. 74, 281300.
Holley, M.L., Mitson, R.B. and Pratt, A.R. (1975) Developments in sector scanning sonar. IERE Conf.
Proc. 32, 13953.
Holliday, D.V. (1977) The use of swimbladder resonance in the sizing of schooled pelagic fish. Rapp.

P.-v. Reun. Cons. Perm. Int. Explor. Mer 170, 1305.


Jolly, G.M. and Hampton, I. (1990) A stratified random transect for acoustic surveys of fish stocks.
Can. J. Fish. Aquat. Sci. 47, 128291.
Kloser, R.J. (1996) Improved precision of acoustic surveys of benthopelagic fish by means of a deeptowed transducer. ICES J. Mar. Sci. 53, 40713.
Knudsen, H. (1990) The Bergen Echo Integrator; an introduction. J. Cons. Int. Explor. Mer 47,
16774.
Lamboeuf, M., Burczynski, J., Bencherifi, M., Chbani, M. and Elminowicz, A. (1983) Evaluation
acoustique de la biomasse de stocks de sardine au Maroc de 1979 a 1981. Combinaison des
`
estimations du sonar et du sondeur vertical. FAO Fish. Rep. 300, 197207.
Levenez, J.J., Gerlotto, F. and Petit, D. (1990) Reaction of tropical coastal pelagic species to artificial

lighting and implications for the assessment of abundance by echo integration. Rapp. P.-v. Reun.
Cons. Int. Explor. Mer 189, 12834.
Love, R.H. (1977) Target strength of an individual fish at any aspect. J. Acoust. Soc. Am. 62,
1397403.
Lytle, D.W. and Maxwell, D.R. (1983) Hydroacoustic assessment in high density fish schools. FAO
Fish. Rep. 300, 15771.
MacCartney, B.S. and Stubbs, A.R. (1971) Measurements of the acoustic target strength of fish in
dorsal aspect, including swimbladder resonance. J. Sound Vib. 15, 397420.
MacLennan, D.N. (1990) Acoustical measurement of fish abundance. J. Acoust. Soc. Am. 87, 115.
MacLennan, D.N. and MacKenzie, I.G. (1988) Precision of acoustic fish stock estimates. Can. J. Fish.
Aquat. Sci. 45, 60516.
MacLennan, D.N. and Simmonds, E.J. (1992) Fisheries Acoustics. London: Chapman & Hall. 336 pp.
MacLennan, D.N., Hollingworth, C.E. and Armstrong, F. (1989) Target strength and the tilt angle
distribution of caged fish. Proceedings of the Institute of Acoustics I.O.A. 11, 1121.
MacLennan, D.N., Magurran, A.E., Pitcher, T.J. and Hollingworth, C.E. (1990) Behavioural

determinants of target strength. Rapp. P.-v. Reun. Cons. Int. Explor. Mer 189, 24553.
Main, J. and Sangster, G.I. (1981) A study of the fish capture process in a bottom trawl by direct
observations from an underwater vehicle. Scott. Fish. Rep. 23, 124.
Marchal, E. and Petitgas, P. (1993) Precision of acoustic fish abundance estimates: separating the
number of schools from the biomass in the schools. Aquat. Liv. Resour. 6, 21119.
Masse, J. and Retiere, N. (1995) Effect of number of transects and identification hauls on acoustic

`
biomass estimates under mixed species conditions. Aquat. Liv. Resour. 8, 1959.
Masse, J., Koutsikopoulos, C. and Patty, W. (1996) The structure and spatial distribution of pelagic

fish schools in multispecies clusters: an acoustic study. ICES J. Mar. Sci. 53, 15560.
Midttun, L. and Hoff, I. (1962) Measurements of the reflection of sound by fish. FiskDir. Skr. Ser.
Havunders. 13, 118.
Misund, O.A. (1993) Dynamics of moving masses: variability in packing density, shape, and size
among herring, sprat and saithe schools. ICES J. Mar. Sci. 50, 14560.
Misund, O.A. (1994) Swimming behaviour of fish schools in connection with capture by purse seine
and pelagic trawl. In Ferno, A. and Olsen, S., eds. Marine Fish Behaviour in Capture and

Underwater acoustics in marine fisheries

31

Abundance Estimation. Oxford: Fishing News Books, pp. 84106.


Misund, O.A. and Aglen, A. (1992) Swimming behaviour of fish schools in the North Sea during
acoustic surveying and pelagic trawl sampling. ICES J. Mar. Sci. 49, 32534.
Misund, O.A., Aglen, A., Beltestad, A.K. and Dalen, J. (1992) Relationships between the geometric
dimensions and biomass of schools. ICES J. Mar. Sci. 49, 30515.
Misund, O.A., Totland, B., Floen, S. and Aglen, A. (1994) Computerbased detection of fish schools by
a multibeam sonar. In Bjrn, L., ed. Proc. 2nd Eur. Conf. Underwater Acoustics, Copenhagen,
Denmark, 48 July. Amsterdam: Elsevier, pp. 81520.
Misund, O.A., Aglen, A. and Frns, E. (1995) Mapping the shape, size and density of fish schools
by echo integration and a high resolution sonar. ICES J. Mar. Sci. 52, 1120.
Misund, O.A., vredal, J.T. and Hafsteinsson, M.T. (1996) Reactions of herring schools to the sound
field of a survey vessel. Aquat. Liv. Resour. 9, 511.
Mitson, R.B. (1983a) Fisheries Sonar. London: Fishing News Books. 287 pp.
Mitson, R.B. (1983b) Acoustic detection and estimation of fish near the sea-bed and surface. FAO
Fish. Rep. 300, 2734.
Mitson, R.B. and Storeton West, T.J. (1971) A transponding acoustic fish tag. Radio Electron. Eng.
41, 4839.

Mitson, R.B. and Holliday, D.V (1990) Future developments in fisheries acoustics. Rapp. P.-v. Reun.
.
Cons. Int. Explor. Mer 189, 8291.

Nakken, O. and Olsen, K. (1977) Target strength measurements of fish. Rapp. P.-v. Reun. Cons. Int.
Explor. Mer 170, 5269.
Neproshin, A.Y. (1979) Behaviour of the pacific mackerel, Pneumatophorus japonicus, when affected
by vessel noise. J. Ichthyol. 18, 6959.
Novarini, J.C. and Bruno, D.R. (1982) Effects of the sub-surface bubble layer on sound propagation.
J. Acoust. Soc. Am. 72, 50114.
Nunnallee, E.P. (1991) An investigation of the avoidance reactions of Pacific whiting (Merluccius
productus) to demersal and midwater trawl gear. ICES CM 1991/B:5, 10 pp.
Ohshimo, S. (1996) Acoustic estimation of biomass and school character of anchovy Engraulis
japonicus in the East China Sea and the Yellow Sea. Fish. Sci. 62, 3449.
Olsen, K. (1979) Observed avoidance behaviour in herring in relation to passage of an echo survey
vessel. ICES CM 1979/B:53, 9 pp.

Olsen, K. (1990) Fish behaviour and acoustic sampling. Rapp. P.-v. Reun. Cons. Int. Explor. Mer 189,
14758.
Olsen, K. and Ahlquist, I. (1989) Target strength of fish at various depth, observed experimentally.
ICES CM 1989/B:35, 8 pp.
Olsen, K., Angell, J. and Lvik, A. (1983a) Quantitative estimations of the influence of fish behaviour
on acoustically determined fish abundance. FAO Fish. Rep. 300, 13949.
Olsen, K., Angell, J., Pettersen, F. and Lvik, A. (1983b) Observed fish reactions to a surveying vessel
with special reference to herring, cod, capelin and polar cod. FAO Fish. Rep. 300, 1318.
Ona, E. (1990) Physiological factors causing natural variations in acoustic target strength of fish.
J. Mar. Biol. Ass. U.K. 70, 10721.
Ona, E. (1994) Recent development of acoustic instrumentation in connection with fish capture and
abundance estimation. In Ferno, A. and Olsen, S., eds. Marine Fish Behaviour in Capture and

Abundance Estimation. Oxford: Fishing News Books, pp. 20016.


Ona, E. and Eger, K. (1989) Fish behaviour and trawl geometry observed with a scanning net-sonde
sonar. In Proc. World Symp. Fish. Gear Fish. Vessel Design, St. Johns, Newfoundland, 1988. St.
Johns, Newfoundland, Canada: Marine Institute, pp. 41216.
Ona, E. and God, O.R. (1990) Fish reactions to trawling noise: the significance for trawl sampling.

Rapp. P.-v. Reun. Cons. Int. Explor. Mer 189, 15966.


Ona, E. and Mitson, R.B. (1996) Acoustic sampling and signal processing near the seabed: the
deadzone revisited. ICES J. Mar. Sci. 53, 67790.

32

Misund

Ona, E. and Toresen, R. (1988a) Avoidance reactions of herring to a survey vessel, studied by
scanning sonar. ICES CM 1988/H:46, 8 pp.
Ona, E. and Toresen, R. (1988b) Reactions of herring to trawling noise. ICES CM 1988/B:36, 8 pp.
Petitgas, P. (1993) Geostatistics for fish stock assessments: a review and an acoustic application. ICES
J. Mar. Sci. 50, 28598.
Petitgas, P. and Levenez, J.J. (1996) Spatial organization of pelagic fish: echogram structure, spatio
temporal condition, and biomass in Senegalese waters. ICES J. Mar. Sci. 53, 14753.
Petitgas, P. and Prampart, A. (1993) EVA: a geostatistical software on IBM-PC for structure
characterization and variance computation. ICES CM 1993/D:65, 55 pp.
Pitcher, T.J. (1983) Heuristic definitions of shoaling behaviour. Anim. Behav. 31, 61113.
Pitcher, T.J., Misund, O.A., Ferno, A., Totland, B. and Melle, V (1996) Adaptive behaviour of herring

.
schools in the Norwegian Sea as revealed by high-resolution sonar. ICES J. Mar. Sci. 53,
44952.
Richards, L.J., Kieser, R., Mulligan, T.J. and Candy, J.R. (1991) Classification of fish assemblages
based echo integration surveys. Can. J. Fish. Aquat. Sci. 48, 126472.
Rose, G.A. (1992) A review of problems and new directions in the application of fisheries acoustics
on the Canadian East Coast. Fish. Res. 14, 10528.
Rose, G.A. (1993) Cod spawning on a migration highway in the northwest Atlantic. Nature 366,
45861.
Rose, G.A. and Leggett, W.C. (1988) Hydroacoustic signal classification of fish schools by species.
Can. J. Fish. Aquat. Sci. 45, 597604.
Rose, G.A., deYong, B. and Colbourne, E.B. (1995) Cod (Gadus morhua L.) migration speeds and
transport relative to currents on the north-east Newfoundland Shelf. ICES J. Mar. Sci. 52,
90314.
Rttingen, I. (1976) On the relation between echo intensity and fish density. FiskDir. Skr. Havunders.
16, 30114.
Rttingen, I., Foote, K.G., Huse, I. and Ona, E. (1994) Acoustic abundance estimation of the
wintering Norwegian spring spawning herring, with emphasis on methodological aspects. ICES
C.M. 1994/(B D G H):1, 17 pp.
Samb, B. and Demarq, H. (1989) Influence of the Senegalese upwelling on the behaviour of pelagic
fish species described by echo-integration. Proc. Inst. Acoust. 11, 10414.
Scalabrin, C. and Lurton, X. (1994) Fish shoals echo amplitude analysis. In Bjrn, L., ed. Proc. 2nd
Eur. Conf. Underwater Acoustics, Copenhagen, Denmark, 48 July. Amsterdam: Elsevier,
pp. 80714.
Scalabrin, C. and Masse, J. (1993) Acoustical detection of the spatial and temporal distribution of fish

shoals in the Bay of Biscay. Aquat. Liv. Resour. 6, 26983.


Scalabrin, C., Diner, N., Weill, A., Hillion, A. and Mouchot, M.-C. (1996) Narrowband acoustic
identification of monospecific fish shoals. ICES J. Mar. Sci. 53, 1818.
Shibata, K. (1971) Experimental measurements of target strength of fish. In Kristjonsson, H., ed.
Modern Fishing Gear of the World, Vol. 2. London: Fishing News (Books), pp. 1048.
Simard, Y., Marcotte, D. and Bourgault, G. (1993) Exploration of geostatistical methods for mapping
and estimating acoustic biomass of pelagic fish in the gulf of St. Lawrence: size of echointegration unit and auxilliary environmental variables. Aquat. Liv. Resour. 6, 18599.
Simmonds, E.J. (1990) Very accurate calibration of a vertical echo sounder: a five-year assessment of

performance and accuracy. Rapp. P.-v. Reun. Cons. Int. Explor. Mer 189, 18391.
Simmonds, E.J. and Armstrong, F. (1990) A wideband echo sounder: measurements on cod, saithe,

herring and mackerel from 27 to 54 kHz. Rapp. P.-v. Reun. Cons. Int. Explor. Mer 189, 3817.
Simmonds, E.J. and Fryer, R.J. (1996) Which are better, random or systematic acoustic surveys? A
simulation using North Sea herring as an example. ICES J. Mar. Sci. 53, 3950.
Simmonds, E.J., Williamson, N., Gerlotto, F. and Aglen, A. (1992) Acoustic survey design and analysis
procedures: a comprehensive review of good practice. ICES Coop. Res. Rep., No. 187, 127 pp.

Underwater acoustics in marine fisheries

33

Simmonds, E.J., Armstrong, F. and Copland, P.J. (1996) Species identification using wideband
backscatter with neural network and discriminant analysis. ICES J. Mar. Sci. 53, 18995.
Smith, G.W., Hawkins, A.D., Urquhart, G.G. and Shearer, W. (1981) Orientation and energetic
efficiency in the offshore movements of returning Atlantic salmon, Salmo salar L. Scot. Fish.
Res. Rep. No. 21. 22 pp.
Smith, P.E. (1970) The horizontal dimensions and abundance of fish schools in the upper mixed layer
as measured by sonar. In Brooke Farquhar, G., ed. Proc. Int. Symp. Biological Sound Scattering
in the Ocean. Washington, DC: Maury Center for Ocean Science, Dep. Navy, pp. 56391.
Smith, P.E. (1977) The effects of internal waves on fish school mapping with sonar in the California

Current area. Rapp. P.-v. Reun. Cons. Int. Explor. Mer 170, 22331.
Soria, M., Freon, P. and Gerlotto, F. (1996) Analysis of vessel influence on spatial behaviour of fish

schools using a multi-beam sonar and consequences for biomass estimates by echo-sounder. ICES
J. Mar. Sci. 53, 4538.
Soule, M., Barange, M. and Hampton, I. (1995) Evidence of bias in estimates of target strength
obtained with a split-beam echo-sounder. ICES J. Mar. Sci. 52, 13944.
Sund, O. (1935) Echo sounding in fishery research. Nature, 135, 935.
Stanley, D.R. and Wilson, C.A. (1996) Abundance of fishes associated with a petroleum platform as
measured with dual-beam hydroacoustics. ICES J. Mar. Sci. 53, 4735.
Strmme, T. and Stersdal, G. (1987) Consistency of acoustic biomass assessments tested by repeated
survey coverage. International Symposium on Fisheries Acoustics, Seattle, Washington, USA,
June 2226, 1987 (Unpublished).
Swartzman, G., Stuetzle, W., Kulman, K. and Powojowski, M. (1994) Relating the distribution of
pollock schools in the Bering Sea to environmental factors. ICES J. Mar. Sci. 51, 48192.
Swartzman, G., Silverman, E. and Williamson, N. (1995) Relating trends in walleye pollock (Theragra
chalcogramma) abundance in the Bering Sea to environmental factors. Can. J. Fish. Aquat. Sci.
52, 36980.
Tameishi, H., Shinomiya, H., Aoki, I. and Sugimoto, T. (1996) Understanding Japanese sardine
migrations using acoustic and other aids. ICES J. Mar. Sci. 53, 16771.
Thorne, R.E. and Thomas, G.L. (1990) Acoustic observations of gas bubble release by Pacific herring
(Clupea harengus pallasi). Can. J. Fish. Aquat. Sci. 47, 19208.
Toresen, R. (1991) Absorption of acoustic energy in dense herring schools studied by the attenuation
in the bottom echo signal. Fish. Res. 10, 31727.
Traynor, J.J. and Williamson, N.J. (1983) Target strength measurements of walleye pollock (Theragra
chalcogramma) and a simulation study of the dual beam method. FAO Fish. Rep. 300, 11224.
Traynor, J.J., Williamson, N.J. and Karp, W.A. (1990) A consideration of the accuracy and precision of

fish abundance estimates derived from echo-integration surveys. Rapp. P.-v. Reun. Cons. Int.
Explor. Mer 189, 10111.
Urick, R.J. (1983) Principles of Underwater Sound, 3rd edn. New York: MacGrawHill. 384 pp.
Urquhart, G.G. and Stewart, P.A.M. (1993) A review of techniques for the observation of fish
behaviour in the sea. ICES Mar. Sci. Symp. 196, 1359.
Voglis, G.M. and Cook, J.C. (1966) Underwater applications of an advanced acoustic scanning
equipment. Ultrasonics 4, 19.
Vilhjalmsson, H., Reynisson, P., Hamre, J. and Rttingen, I. (1983) Acoustic abundance estimation of

the Icelandic stock of capelin 19781982. FAO Fish. Rep. 300, 20816.
Walsh, M., Reid, D.G. and Turrell, W.R. (1995) Understanding mackerel migration off Scotland:
tracking with echosounders and commercial data, and including environmental correlates and
behaviour. ICES J. Mar. Sci. 52, 92540.
Walsh, S. (1991) Diel variation in availability of and vulnerability of fish to a survey trawl. J. Appl.
Ichthyol. 7, 14759.
Weill, A., Scalabrin, C. and Diner, N. (1993) MOVIES-B: an acoustic detection description software.
Application to shoal species classification. Aquat. Liv. Resour. 6, 25567.

34

Misund

Weston, D.E. and Andrews, H.W. (1990) Monthly estimates of fish numbers using long-range sonar.
J. Cons. Int. Explor. Mer 47, 10411.
Williamson, N.J. and Traynor, J.J. (1996) Application of a one-dimensional geostatistical procedure to
fisheries acoustic surveys of Alaskan pollock. ICES J. Mar. Sci. 53, 4238.
Zakharia, M.E., Magand, F., Hetroit, F. and Diner, N. (1996) Wideband sounder for fish species
identification at sea. ICES J. Mar. Sci. 53, 2038.

Accepted 16 December 1996

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

You might also like