I CES J ournal of MarineScience, 53: 513516.

1996
VIII. Conclusion
Fisheriesandplanktonacoustics: past, present, andfuture
David N. MacLennan and D. V. Holliday
I ntroduction
The success of the Aberdeen Symposiumin J une 1995
hasdemonstratedastrongandcontinuinginterest inthe
subject of sheries and plankton acoustics, as do the
collected scientic papers in this volume. The Sympo-
siumwas also an opportunity to reect on theachieve-
ments of the past, to review current problems, and to
think about the future. This note is a short overview
from our personal standpoints, based on the intro-
ductory and concluding presentations we gave at the
Symposium.
A brief history
The story of hydroacoustics goes back more than 50
years to thepioneering work of Oscar Sund in Norway
and RonnieBallsin England. Theearliest echosounders
wereverylimited theycouldshowonlythepresenceor
absenceof sh. But what wecall themodern era begins
in the 1960s when echo counters and echo integrators
wererst used to determinetheabundanceof sh.
A good way to examineacoustical developmentsover
the past 30 years is through the proceedings of major
conferencesinthat period. Theseriesbeganwiththetwo
Bergenconferencesin1973and1982. Next therewasthe
Seattlemeeting in 1987 and now theAberdeen Sympo-
sium in 1995. Looking through the papers from these
symposia, it isinterestingto seehowthequestionsbeing
addressed havechanged.
At thetimeof the1973conferencein Bergen, theidea
of acoustic surveying of sh populations was very
new, and acoustic surveys as such did not receive
much attention. There were several papers on acoustic
methods in sh behaviour research, for exampletheuse
of acoustic tags attached to sh to study their move-
ments. Probablythemost important contributionsat the
1973 Bergen conference were those on the target
strengthof sh, asubject whichwasnot well understood
at thetime. I nparticular, therewastheseminal paper by
NakkenandOlsenabout their experimental work onthe
target strength of sh. Although better techniques are
now available, which can be used with live free-
swimming sh, the Nakken and Olsen paper continues
to appear in referencelists as a corepaper in this eld.
The technique of echo integration depends on the
fundamental assumption that acoustic scatteringby sh
is a linear process. That is to say, it is assumed that the
echo energy frommultipletargetsisthesumof theecho
energies which would be received from each target in
isolation. During the 1970s, there was erce argument
on thevalidity of thelinearity principle. Someheld that
the distribution of sh in schools was not suciently
randomfor linearity to apply, in which casethemethod
of echo integration would not be valid. Clearly, it was
important that such a fundamental assumption should
be tested by experiment, but early attempts to do this
failed, mainly because passive targets were used which
could not replicate the conditions applying in live sh
schools. Then Foote conducted a denitive experiment
with live sh, showing that linearity did apply to sh
densities typical of those encountered on acoustic sur-
veys. Foote reported his linearity experiments to the
second Bergen conferencein 1982, and his contribution
was certainly a highlight of that meeting.
Another important advance around 1980 was the
development of accurate methods for calibrating echo-
sounders and sonars. Before then, echosounders were
calibrated using various methods, such as reference
hydrophones, but accuracy was poor and calibration
errors could be4050%or even worse. However, it was
foundthat metal sphereswithknownacousticproperties
could be used as reference targets to produce a well-
dened echo which is measured to calibrate the output
of the echosounder. This is called the standard target
method and allows calibrations to be performed to an
accuracy of about 5%. There are many other errors in
acousticabundanceestimation much larger than this, so
it has not been necessary to look for further improve-
ments in the calibration technique, at least for current
applications. Fromthetimeof theBergen conferencein
1982, thecalibrationof sonarsandechosounderswasno
longer themajor problemin shery acousticsthat it had
been previously.
The Seattle conference was only ve years later, but
there had been rapid development in the meantime.
Perhaps the most important advance in the 1980s was
the introduction of dual-beam and split-beam echo-
sounders. These instruments give direct measurements
of thetarget strength of wild sh in their natural state,
theso-called direct in situ method.
I t became clear from the discussions in Seattle that
there was need for new theoretical approaches to the
interpretation of acoustic data from dual-beam and
10543139/96/020513+04 $18.00/0 1996 I nternational Council for theExploration of theSea

b
y

g
u
e
s
t

o
n

S
e
p
t
e
m
b
e
r

6
,

2
0
1
2
h
t
t
p
:
/
/
i
c
e
s
j
m
s
.
o
x
f
o
r
d
j
o
u
r
n
a
l
s
.
o
r
g
/
D
o
w
n
l
o
a
d
e
d

f
r
o
m

split-beam instruments. I t was seen that a statistical
approachwasnecessarytotakeaccount of thestochastic
nature of target strength, which can vary over a large
rangeof valueseven for thesamesizeof sh and, in any
case, observed sh populations are rarely of constant
sizeso thesizedistribution of theensonied sh has to
beconsidered asan additional stochastic feature. Kieser
and Ehrenberg made an important contribution to the
theoretical debate with their unbiased stochastic echo-
counting model which was presented at the Seattle
meeting. They showed that if the statistics of target
strength and sh-size distributions were ignored, then
sh-abundance estimates obtained with simple non-
stochastic models could beseriously in error.
TheSeattlemeeting was also notablefor theinterest
shown in the application of acoustic methods to new
and more dicult problems in marine and freshwater
science. Studies of zooplankton and freshwater sh
populationsreceivedmuchattentionwhichhadnot been
the case at the earlier conferences. Exciting new devel-
opments were reported, notably the use of wideband
echosounders to identify sh targets fromtheir acoustic
signature, andthegeostatistical approachto theanalysis
of survey results was mentioned for therst time.
Thepresent
The last few years have continued to see innovative
developments in acoustic technology, in signal process-
ing, in methods for designing acoustic surveys and for
the analysis of survey results. Fish behaviour research
has greatly improved our understanding of target
strength and why it is a highly variable parameter.
Powerful statistical techniques have been applied to
improvethereliability of abundanceestimates obtained
fromacoustic measurements.
As a result of these developments, acoustic methods
have become well established in many areas of marine
and freshwater investigations. Moreover, it is fair to
say that hydroacoustics has matured to become a
respectable measurement technique and not the black
art that it once was. I t is now realistic to expect that
acoustic measurements will be presented with objective
condence limits, and not simply as a number whose
signicancenonebut theacoustician can understand.
All this was clearly reected in presentations at the
Aberdeen Symposium. We are no longer consumed by
themes that address instrumentation problems, ques-
tions about what acoustical measurements mean, and
how they relate to sh biomass estimated by other
methods. The emphasis now is on survey strategies
and sophisticated processing techniques for extracting
information on aquatic animals and their environment.
We have seen presentations on scales from individual
plankton to the structure of large sh aggregations.
Much of this work has an ecosystem dimension, with
ancillary dataon thephysical environment and thefood
web being collected along with the usual acoustical
records and sh sampling. This conrms the trend
towards interdisciplinary studies in oceanography. I t is
also consistent with the broader view within the I CES
community that understanding the interaction between
physical oceanographyandthefoodwebisimportant to
thelarger problemof sustainablesheries management.
Plankton acoustics
Although plankton have been studied acoustically for
many years, thegrowingimportanceof this eld is now
evident. There were a large number of contributions
covering both small zooplankton and micronekton,
indicating a healthy and continuing interest in this
discipline.
Three notable themes came out of the discussions in
Aberdeen. Firstly, there is sucient understanding of
scatteringfromplanktontoconcludethat thetraditional
regressions used in sheries acoustics are not adequate
descriptors of the relationships between the acoustical
reectivity and theabundance, size, species, genera, and
behaviour of plankton. Secondly, acoustic instruments
must sample plankton with high resolution in the fre-
quency domain and over a wide bandwidth to give
useful data. Thirdly, it isclear that thecomplexscattered
sound eld of plankton contains information which, if
properly understood, could lead to useful estimates of
abundance and other biophysical parameters. I n some
instances at least, the classication of targets could be
much improved through the combination of acoustical
data with information on the environment and animal
behaviour.
TheAcoustic Doppler Current Proler (ADCP)
An important contribution conrmed the view that
exceptional care is needed when correcting for instru-
ment artifacts and variability, in thecourseof convert-
ing ADCP backscattering results to distributions of
biomass. Several investigatorshavesuggestedthat useful
data on thedistribution of plankton, micronekton, and
mesopelagicshesmaybeobtainedwiththeADCP, but
theeort to makethenecessary corrections is far from
trivial. Theanalysis of ADCP backscatteringdata must
takeaccount of thefact that theobservations represent
combined eects of changes in abundanceand changes
in size. I t remains to beseen whether this complication
in interpretingADCP datawill allowuseful estimatesof
plankton or micronekton biomass.
Applications in limnology
I t isrefreshingto notethat theuseof acoustical methods
in limnology is maturing rapidly. Presentations at the
514 D. N. MacLennan and D. V. Holliday

b
y

g
u
e
s
t

o
n

S
e
p
t
e
m
b
e
r

6
,

2
0
1
2
h
t
t
p
:
/
/
i
c
e
s
j
m
s
.
o
x
f
o
r
d
j
o
u
r
n
a
l
s
.
o
r
g
/
D
o
w
n
l
o
a
d
e
d

f
r
o
m

Symposium have described sophisticated ecosystem
studies which apply acoustics as an established tool
rather than an overwhelmingeort in itself. As this is a
relatively youngdiscipline, it is encouragingto nd that
most of the active participants are not still struggling
with the kind of instrumentation issues that so domi-
nated marine sheries acoustics in the early years. The
useof acoustics in limnology is potentially high prole,
sincetheapplication to recreational as well as commer-
cial sheries could havea long-termpositiveimpact on
the wider community. There is an important public
relations aspect in our contribution to the eective
assessment and protection of limnological resources.
Vessel noise
The eect of the noise radiated by sheries research
vessels on survey results has recently been evaluated by
an I CES working group. The conclusion is clear the
designers of shery research vessels must give serious
consideration to theradiated noiseproblem. Theunre-
solved issue is whether the scientic recommendations
will be given due weight by those responsible for the
procurement and overhaul of research vessels.
Thefuture
There are some important topics where outstanding
problems remain to be resolved. Further research on
these issues is required to develop the full potential of
sheries and plankton acoustics.
Target strength
The acoustical reectivity of sh and plankton is a
subject of continuing interest. For the purpose of bio-
massassessment, thereisnowreasonableunderstanding
of what target strength means and how it is estimated.
Careful experiments to determine target strength allow
reasonable condence in the results of many acoustic
surveys, particularly those for common, monospecic
stocks. TheI CESFAST study group on target-strength
methodology is makinga valuablecontribution, but we
believe their work is only an early step on a long road
for the broad eld of bioacoustics. Many problems
remain to be resolved concerning the dependence of
target strength on sh behaviour and the environment.
I n particular, the issue of single versus multiple targets
recognition is not yet resolved. This issue is receiving
badly needed attention and progress is being made, but
it isclear that additional work andeducationof theuser
community is required.
I t is not a practical proposition to measure the
acoustical reectivity of all species at all frequencies,
in all physiological conditions, environmental cir-
cumstances, and behaviours which might occur in the
eld. A complete solution depends on having a predic-
tivecapability that is soundly based on an understand-
ing of rst principles. However, present understanding
of echo formation fromrst principles is insucient to
predict target strengthfor anybut averylimitednumber
of species of plankton and nekton. Wemust struggleto
achievesuch a capability based on animal morphology,
physiology, behaviour, and the physics of echo forma-
tion. To do thisin aproper and useful way isoneof the
most critical tasksin sheriesand plankton acoustics. I t
will requireinnovativethought, mathematical modelling
at a level of detail we have not yet encountered in our
eld, and newmethodology for validation of modelling
results. Thevariedskillsrequiredfor thistask meanthat
a cooperativeeort is indicated.
Species identication
This is the grand challenge of sheries and plankton
acoustics. I t will not be accomplished with acoustics
alone, but it cannot beaccomplished without acoustics.
All available information must be integrated in new
analytical procedures which will make probabilistic
statements regarding the specic cause of the small,
strangelyshaped, intriguingmarksseenonechosounders
and sonars.
I n spite of the availability of an impressive suite of
analytical tools such as neural net processing, expert
systems, and new methods of discriminant analysis, we
believe that these data-processing aids cannot success-
fully resolve the problem of acoustically aided species
identication without being fed clues with better dis-
crimination capabilities than areavailableat present. I n
general, existingsurveytechniquesdonot measuremany
of the features that are necessary for accurate target
classication. I n the most general case, it may well be
impractical to extract good classication clues froman
acoustical sensor that is optimized for biomass assess-
ment. A well-established theoremin information theory
holds that the information-carrying capacity of a com-
munications channel depends on its bandwidth. Thus,
thetrend towards theuseof wideband systems is likely
to bea fruitful approach.
I ncreasingthebandwidth in acoustical signal process-
inghasthepotential toachievegreater overall resolution
with an acoustical sensor. I n essence, oneimproves the
ability to distinguish the individuals in a school or
aggregation and perhaps even the component parts of
individual animals. I nformation so derived provides
detail for examining structure, shape, ping-to-ping
motion, and ne-scale distribution. I n addition to
increasing range resolution, there are other dimensions
to be explored, such as spectral analysis of echoes
using the Doppler eect to reveal target motion and
behaviour.
515 Fisheries and plankton acoustics: past, present, and future

b
y

g
u
e
s
t

o
n

S
e
p
t
e
m
b
e
r

6
,

2
0
1
2
h
t
t
p
:
/
/
i
c
e
s
j
m
s
.
o
x
f
o
r
d
j
o
u
r
n
a
l
s
.
o
r
g
/
D
o
w
n
l
o
a
d
e
d

f
r
o
m

This is only a partial list of theclassication clues
that shouldeventuallyrealizethefull potential of acous-
tic methods. Researchers must not despair of thesome-
times limited applicability and results of early attempts
to usediscrimination techniques transferred fromother
disciplines. Their eventual successdependsontheability
to extract well-conceived, ecient, acoustically derived
discriminantsfor classication. Without attentiontothis
area, it will begarbagein, garbageout. Weareonlyat
thebeginning of a long road towards theidentication
of sh and plankton by acoustical means. With good
discriminants, we should eventually achieve the ability
to placeaccurateprobabilities and condencelimits on
theidentication of species.
Computing and instrumentation
Many of these developments have depended on the
availability of cheap computers which are vastly more
powerful than thoseon themarket just a fewyears ago.
The processing power of modern computers has had a
hugeinuenceon scientic research in general, and the
benets in hydroacoustics have been no less than in
other elds. Computing power has led to sonars and
echosounders which display more information about
what is going on, yet modern acoustic instruments are
relatively simpleto operatewith theassistanceof micro-
processors which provide menu displays, help screens,
and all the rest of it. The data collected fromacoustic
surveyscanbestoredingigabytequantities, whichallow
post-survey processing of the results to an extent that
wasimpossiblenot so verylongago. Ontheother hand,
therearerisksin lettingcomputersdo moreof thework
without human intervention. Computers are only as
good as the programs which control them, and the
programs are only as good as the people who wrote
them. Mistakes in computer softwarecan and do occur
and theconsequences can bedisastrous to say theleast.
I t is essential for operational procedures to include
checks with human intervention to ensure the results
comingfromtheautomatic data collection and analysis
aresensible. Thetimemay comewhen acoustic surveys
can bedonealmost entirely under machinecontrol, but
we suggest that that is a development for the long
distant future.
The earlier comments on the adequacy of current
methods for calibration apply only to conventional
surveyinstrumentation. Weanticipatethat newresearch
techniques will bring new challenges in calibration.
Systemsdesigned to optimally addressproblemssuch as
near-boundarydetectionandtheclassicationof targets
will not necessarily have the same characteristics as
thosewerely upon today. Weenvisagedevelopments in
thenext decadebased on wideband or coded transmis-
sions, parametric arrays, and the use of frequencies
much lower and much higher than at present. With
further advancesinelectronicsandcomputer processing,
new kinds of multi-frequency and multi-beam systems
will becomeavailablefor routinework at an aordable
cost. The measurement standards and parameter de-
nitions applicable to underwater acoustics may need
renement and extension as more sophisticated and
specialized systems comeinto general use.
Education in sheries acoustics thenext
generation
Thereisan increasingdemand for an acoustical compo-
nent in ecosystem studies involving sh and plankton.
This suggests that educational programmes in the con-
duct and practiceof our specialized disciplines must be
enhanced. However, it is essential for educators to
consider what the actual carrying capacity of the
discipline really is, before training large numbers of
peoplefor jobs that may not materialize.
I n addition to the personnel needed for traditional
resource assessment surveys, it will be necessary to
recruit innovativephysicists, acousticians, and engineers
into the research and development community, to sup-
port the evolution of new instrumentation for use in
sheries and plankton investigations. Theattendanceof
many representatives from academia at the Aberdeen
Symposium is an encouraging sign of a developing
interest in thebasic scienceof our discipline.
Fisheries and plankton acoustics arecharacterized by
anintelligent, careful, innovative, andsophisticateduser
community which provides a valuable service to I CES
and other customers. This community must support
basic research on the conception and development of
newtechniquesfor thefuture. Thekeytoprogressisnew
blood with newideas and approaches.
TheAberdeen Symposiumhas attracted a signicant
number of fresh, new, youngfaces. They havepresented
valuablescienticcontributions. Theywill formthenext
generation of sheries and plankton acousticians. I t is
likelythat manyof our generationwill soonbeseekinga
more peaceful life, not including long months at sea in
rough weather. But it remainsour responsibility to train
the young scientists, and to encourage themto pursue
sheriesacousticsandrelatedsubjectswithnewperspec-
tives. I t couldbethat themost important outcomeof the
Aberdeen Symposium is the extent to which it has
contributed to that end.
D. N. MacLennan: SOAFD MarineLaboratory, PO Box
101, Victoria Road, Aberdeen, AB9 8DB, Scotland, UK.
D. V. Holliday: Tracor AppliedSciences, Suite102, 4669
Murphy Canyon Road, San Diego, CA 92123, USA.
Correspondence to MacLennan [tel: +44 1224 876544,
fax: +44 1224 295511].
516 D. N. MacLennan and D. V. Holliday

b
y

g
u
e
s
t

o
n

S
e
p
t
e
m
b
e
r

6
,

2
0
1
2
h
t
t
p
:
/
/
i
c
e
s
j
m
s
.
o
x
f
o
r
d
j
o
u
r
n
a
l
s
.
o
r
g
/
D
o
w
n
l
o
a
d
e
d

f
r
o
m